AI/ML PhD Intern - Summer 2026
About the job
- Current PhD student studying AI/ML at a U.S. accredited university
- Deep understanding of machine learning, including deep learning algorithms, model architectures and optimization
- Baseline understanding of building and deploying production-grade AI/ML systems
- Knowledge of the full ML lifecycle: data curation, training, deployment, monitoring
- Desire to bridge research and application—bringing new AI ideas from concept to impactful deployment.
- Conduct original AI/ML research in search, recommendations, natural language processing, agentic workflows and more
- Prototype and productionize AI systems that feel intuitive but do a lot under the hood
- Tackle novel UX problems at the intersection of AI, analytics, and data apps
- Contribute to the development of scalable AI/ML infrastructure for experimentation, training, and evaluation of large-scale models.
- You have built agents that can plan, reason, and use tools
- You’ve fine-tuned or trained foundation models (language or multimodal) for novel domains
- You know your way around cloud infrastructure (AWS, GCP, Azure)
- You’ve worked in a fast-moving startup or high-growth environment
About the Role
At Sigma, we’re not just adding AI—we’re building the future of how people work with data. Our platform already lets users explore billions of rows of data in seconds with a spreadsheet-like interface, analyze and present their data in workbooks, and build data apps and workflows. Now we’re pushing further, applying AI to reshape how people build in Sigma, discover insights, and make smarter decisions—fast. You can learn more about our vision for AI/ML at Sigma in our recent blog post.
That’s where you come in. As an AI/ML PhD Intern, you’ll join a growing team focused on building the AI foundation that will power Sigma for the future. Your work will become an integral part of the workflow for the thousands of enterprises that run on Sigma. We're looking for PhD students that are passionate about the intersection of AI, business, and data.
What You’ll Do
- Conduct original AI/ML research in search, recommendations, natural language processing, agentic workflows and more
- Prototype and productionize AI systems that feel intuitive but do a lot under the hood
- Tackle novel UX problems at the intersection of AI, analytics, and data apps
- Contribute to the development of scalable AI/ML infrastructure for experimentation, training, and evaluation of large-scale models.
What You Bring
- Current PhD student studying AI/ML at a U.S. accredited university
- Deep understanding of machine learning, including deep learning algorithms, model architectures and optimization
- Baseline understanding of building and deploying production-grade AI/ML systems
- Knowledge of the full ML lifecycle: data curation, training, deployment, monitoring
- Desire to bridge research and application—bringing new AI ideas from concept to impactful deployment.
Bonus Points (or skills you’ll build here)
- You've built agents that can plan, reason, and use tools
- You’ve fine-tuned or trained foundation models (language or multimodal) for novel domains
- You know your way around cloud infrastructure (AWS, GCP, Azure)
- You’ve worked in a fast-moving startup or high-growth environment
Additional Job details
The hourly rate range for this position is $80-$90 an hour
Compensation may vary outside of this range depending on a number of factors, including a candidate’s qualifications, skills, competencies and experience. Base pay is one part of the Total Package that is provided to compensate and recognize employees for their work at Sigma Computing.
About us:
Sigma is the only cloud analytics and business intelligence tool empowering business teams to break free from the confines of the dashboard, explore data for themselves, and make better, faster decisions. The award-winning software was built to capitalize on the performance power of cloud data warehouses to combine data sources and analyze billions of rows of data instantly via an intuitive, spreadsheet-like interface – no coding required.
Since launching with its unique interface, Sigma Computing has added features such as collaboration tools and embedded analytics capabilities. The most recent product launch included a set of AI tools such as forecasting capabilities, an AI copilot and a notebook interface for users who prefer a code-first environment.
Sigma announced its $200M in Series D financing in May 2024, to continue transforming BI through its innovations in AI infrastructure, data application development, enterprise-wide collaboration, and business user adoption. Spark Capital and Avenir Growth Capital co-led the Series D funding round, with additional participation from a group of past investors including Snowflake Ventures and Sutter Hill Ventures.The Series D funding, raised at a valuation 60% higher than the company’s Series C round three years ago, promises to further accelerate Sigma’s growth.
Sigma Computing is an equal opportunity employer. We are committed to building a smart and strong team regardless of race, color, ancestry, religion, sex, national origin, sexual orientation, age, citizenship, marital status, disability, gender, gender identity or expression, or veteran status. We look forward to learning how your experience can enable all of us to grow.
Note: We have an in-office work environment in all our offices in SF, NYC, and London.
About us:
Sigma is the AI apps and analytics platform connected to the cloud data warehouse. Using Sigma, business and technical teams can build intelligent, production-ready AI apps that accelerate and automate operational workflows. Sigma provides a spreadsheet interface, SQL and Python editors, visual builders, and native AI to help teams turn live data into interactive applications, analysis, reports, and embedded experiences.
Sigma announced its $200M in Series D financing in May 2024, to continue transforming BI through its innovations in AI infrastructure, data application development, enterprise-wide collaboration, and business user adoption. Spark Capital and Avenir Growth Capital co-led the Series D funding round, with additional participation from a group of past investors including Snowflake Ventures and Sutter Hill Ventures.The Series D funding, raised at a valuation 60% higher than the company’s Series C round three years ago, promises to further accelerate Sigma’s growth.
Come join us!
Benefits For Our Full-Time Employees:
- Equity
- Generous health benefits
- Flexible time off policy. Take the time off you need!
- Paid bonding time for all new parents
- Traditional and Roth 401k
- Commuter and FSA benefits
- Lunch Program
- Dog friendly office
Sigma Computing is an equal opportunity employer. We are committed to building a smart and strong team regardless of race, color, ancestry, religion, sex, national origin, sexual orientation, age, citizenship, marital status, disability, gender, gender identity or expression, or veteran status. We look forward to learning how your experience can enable all of us to grow.
Note: We have an in-office work environment in all our offices in SF, NYC, and London.
Our Privacy Practices
When you submit a job application on this site, Sigma processes your personal data for the purposes of evaluating your candidacy for employment at Sigma and as otherwise needed throughout the recruitment and hiring process. Please review Sigma’s Candidate Privacy Notice for more details. Please note that your personal data may be transferred to a country other than the one in which it was provided (including to USA, the UK, and Canada).
Sigma’s use of AI
This hiring process utilizes artificial intelligence tools to assist in candidate screening and assessment. Our AI tools are designed to complement, not replace, human decision-making.
About the company
Sigma Computing builds cloud analytics software for business and data teams that need to explore company data and turn findings into operational workflows. Its workbooks combine a spreadsheet interface with charts and dashboards, so teams work with live warehouse data without downloading separate copies. Users build applications for forecasting, approvals and other business processes, while administrators control access through the permissions of the connected data platform. The company sells subscriptions with licenses matched to whether users view reports, contribute data, analyze information or build applications. Founded in 2014, Sigma Computing is headquartered in San Francisco.
Company Size
501-1,000
Company Stage
Series D
Total Funding
$581.3M
Headquarters
San Francisco, CA
Founded
2014
Get referred to Sigma Computing
See people who can refer or advise you
Simplify's Take
What believers are saying
- Demand is compounding: Sigma reported $200 million in annual recurring revenue, doubling year-over-year, in April 2026.
- Equity benchmark doubled: Sigma raised $80 million at a $3 billion valuation in May 2026.
- Broad rollout faces less friction: Sigma gives dashboard consumers free View licenses.
What critics are saying
- Customer bills extend beyond licenses: Sigma warns automatic refreshes create significant warehouse costs.
- Automated reports risk stale results: Sigma scheduled exports reuse the last Python output instead of rerunning code.
- Office work and travel are explicit: Sigma's listed San Francisco backend engineer role expects 25% travel.
What makes Sigma Computing unique
- Sigma captures decisions alongside analysis: input tables write user entries into warehouse schemas with separate edit logs.
- Teams build together: Sigma supports simultaneous editing of the same live workbook without locking files.
- Metrics travel across tools: Sigma accepts dbt, Snowflake and Databricks semantic models without redefining them.
Help us improve and share your feedback! Did you find this helpful?
Benefits
- Competitive pay - Looking for a great salary and solid stock options? You've come to the right place.
- Flexible schedule - Do the work you need to get done in the time you have to get it done
- Amazing benefits - Medical, dental, vision, 401k, FSA, commuter… we've got you covered. Literally.
- Flexible vacation - At Sigma, we work to live, not live to work. So go on, book that dream vacation.
- Health & wellness - A healthy body supports a healthy mind, so we partner with Crunch Fitness and CorePower.
- Family-friendly - From flexible scheduling to parental leave to kids' birthdays off, we support Sigma families.
Growth & Insights and Company News
6 month growth
0%1 year growth
2%2 year growth
3%How Sigma Computing built a game storefront in Sigma with payments and multiplayer mode. Tino Capone App Engineer Jared Avrach App Engineer October 27, 2026 Most AI apps built in Sigma are made to solve specific business problems. Someone has a workflow that needs fixing, or a report that needs creating, and Sigma makes it possible to build a tool for those exact needs and specifications. But at the annual Sigma Hackathon this August, Sigma Computing weren't trying to solve a problem. Sigma Computing were trying to create something that hasn't been done before. At this year's Hackathon, Sigma Computing joined forces to test the limits of building in Sigma. Sigma Computing landed on the idea of building a game storefront, where users could browse a library of games, buy them, and play them without ever leaving Sigma. Using Sigma's plugin and repeater elements, Sigma Computing created a platform where users could browse games, buy them with real Stripe transactions, and play against each other in real time, all without ever leaving Sigma. Sigma Computing called its app "SigmaVerse." The hackathon judges voted it "Best Application," a win that came down to two achievements: * Getting Stripe payments working inside Sigma * Getting real-time multiplayer working through its writeback capability. It was a great experiment to explore the art of the possible with Sigma, and to this day its colleagues are using SigmaVerse to play each other in games like Sigmopoly. Read on to see how Sigma Computing built the app, and the Sigma tricks Sigma Computing learned along the way. How Sigma Computing built SigmaVerse, a gaming storefront in Sigma. The easiest way to explain what Sigma Computing built is to compare it to Steam. Steam is a platform where you can buy individual games made by different studios, load money onto an account, and then play those games, often online against other people. Sigma Computing built the same basic idea, but inside Sigma, and with its own games instead of licensed ones. SigmaVerse is a storefront. You browse a library of games, each tagged with a genre and a difficulty level, and some running a sale with a visible discount. But, instead of buying titles made by outside studios, you are buying familiar games Sigma Computing created ourselves. Sigma Computing built all these games in React with Claude Code, then hosted them for free on Netlify. From there, Sigma Computing relied on Sigma's plugin element to surface these games directly into the app. That same plugin also lets user actions update Input Tables, which is how Sigma Computing save each game's progress and host multiplayer sessions. To simulate the act of buying a game, the user needs funds in their account. So, Sigma Computing opened a Stripe sandbox account and built a Netlify site with an HTML frontend that calls the Stripe API. Then, similar to how Sigma Computing embedded the games, Sigma Computing built a Sigma plugin that links to that Netlify site. So now, when a user adds funds, they enter a fake credit card on the Netlify site, which simulates a real payment through the Stripe sandbox. The plugin then sends a trigger back to Sigma, which runs an action sequence that updates an Input Table with an "Update Row" action, adding the funds to the user's in-game wallet. The balance updates immediately, and the user can then purchase a game and add it to their library. From there, they launch it and play, either on their own or against someone else in real time, all within Sigma. Rewriting what writeback can do. SigmaVerse wouldn't be possible without data writeback. In Sigma, writeback is normally used to update or insert data into the warehouse through an Input Table. Someone edits a value in Sigma, and it saves back to the source. This time, though, Sigma Computing used writeback to let two people communicate with each other inside the same app, in real time, with an Input Table as the element both of them are reading from and writing to. That's really the trick behind its multiplayer. When two people are in the same game, they're both connected to the same Input Table, and that table is acting like a server. Every move one player makes writes back to that table, and the other player sees it update. This is what lets two users sit down and play a game like Sigmopoly. Each player logs into Sigma, navigates to the SigmaVerse app, and clicks on the option to play Sigmopoly. They see a live board built entirely from Sigma elements, as the game itself is a Sigma plugin that acts as a passthrough to a Netlify site, which displays the game's frontend. Each time a player takes an action, like rolling, buying a property, or making a trade, the plugin calls an action sequence that inserts a new game state row into an Input Table. Every player in the game can see that same Input Table, so as soon as a new row lands, everyone's screen updates to reflect the latest state. That game state row carries a JSON payload with everything needed to reconstruct the game in progress: each player's balance, properties owned, trades sent, and more. Every action, every roll, every trade, gets written back to the Input Table as it happens, so both players are always looking at the same live state. The leaderboard next to the game works the same way, where scores get written back and stay updated for anyone checking in. What's useful here isn't limited to games. Any time you need two or more people acting on the same live data inside an app, this pattern applies. You can use writeback to power real-time chat, or enable turn-based collaboration where each person's progress needs to be visible to the others as it happens. Bringing Stripe payments into Sigma. Once Sigma Computing had the multiplayer dimension sorted, Sigma Computing needed to figure out a way for its app to handle real transactions. To make the storefront work, Sigma Computing used a plugin to embed Stripe's API and Checkout UI, then built its own HTML site around it. Sigma plugins are an easy way to embed a specific piece of functionality into an app, like a calendar, a map, or a custom chart. When a SigmaVerse user adds money to their wallet using a fake credit card, the plugin notifies the Stripe API to add funds to the sandbox account. At the same time, the plugin notifies Sigma to run the action sequence that updates the user's balance in the app. It's not a mockup of a payment flow. It's Stripe, running inside a plugin, inside Sigma. In the SigmaVerse app, users can top up their wallets through the Stripe flow, update their balance, and complete a purchase. The game lands in their library, ready to play. Sigma Computing paired that payments plugin with a Sigma repeater element to lay out the storefront, games, prices, sale tags, and difficulty levels, all in a grid that behaves less like a Sigma workbook and more like a shopping app. That pairing, a real transactional backend plus a marketplace-style layout, is what gives the whole thing its shopping-hub feel. Strip away the games, and what's left is a pattern that applies to a lot of real business problems. Any app that needs to track transactional or payment data, an expensing tool, a billing dashboard, or anything with a Stripe account behind it, can use this same setup. And because the transactional data is flowing through Sigma, it's available for analysis the same way any other data in your warehouse is. The marketplace layout on top of it works for more than games, too. Think of anything with a DoorDash or Airbnb feel that features a catalog of items people can browse and buy, with real transactional analytics sitting right underneath. That's what Sigma Computing created inside Sigma. The art of the possible, built in Sigma. By the time Sigma Computing wrapped up, Sigma Computing had a working payments flow, live multiplayer, and a browsable storefront. All of it lived inside Sigma. Sigma Computing built its app around a game because that's what a hackathon gives you room to do. The same tools are sitting there for whatever comes next, whether that's a payments flow, a live collaboration tool, or something Sigma Computing haven't thought to try yet. Start building your own Sigma app: * Explore popular apps and build for free in Sigma Public * Request a Sigma demo * Start a Sigma free trial
The Sigma plugin is live for ChatGPT Work and Codex. September 10, 2026 Sigma Computing is excited to share that the Sigma plugin is now available for ChatGPT Work and Codex, and is part of OpenAI's new Data agent for ChatGPT Work. You can now use the Data agent via ChatGPT to receive answers from Sigma in response to data questions, or even build interactive dashboards in Sigma from a prompt in the ChatGPT interface. This update builds on the existing Sigma plugin, which allows users to query, search, and explore data from Sigma - with the benefit of additional context and governance from Sigma - directly in ChatGPT Work conversations. The updated plugin adds the ability to build a fully interactive dashboard in Sigma straight from a ChatGPT prompt. This means employees can drill deeper into governed data, design and publish a dashboard, and easily share insights with teams. The plugin inherits Sigma's governance and permission settings, so you never have to worry about data access. The Sigma plugin is available today for ChatGPT Work, and Codex to all enterprise OpenAI customers with Sigma Act, Analyze, or Build licenses. The Sigma plugin is also available via the Data agent in the ChatGPT Work Marketplace. Now everyone can put data to work. ChatGPT revolutionized how business teams interact with data, and lowered the bar to performing advanced data analysis. At the same time, it's crucial for every business to make sure that data analysis remains secure, governed, and accurate - and that's exactly what ChatGPT Work and Sigma provide together. It starts with curiosity. Anyone with a paid Sigma license can ask a data question in ChatGPT Work or Codex, and get an answer built from the metrics and workbooks their company already trusts. With another prompt they can turn that answer into an interactive dashboard, complete with custom branding and additional controls to filter, publish, and share the dashboard. This workflow covers the everyday requests that used to start with a ticket: * A product manager preparing a product adoption review * An operations leader analyzing revenue, churn, and net revenue retention * A regional sales manager standing up a live view of quota attainment by rep With the Sigma plugin in ChatGPT, building those dashboards is as easy as hitting "enter." "Connecting Sigma to the Data agent in ChatGPT Work gives teams a conversational way to build on their existing metrics and workbooks. That helps more people understand what's happening in their business and decide where to focus next." - Arpan Shah, General Manager, Technology Vertical, OpenAI How the Sigma plugin in ChatGPT builds a customized, interactive dashboard. The whole flow runs inside the ChatGPT Work or Codex interface, from the first question to a published dashboard in Sigma. 1. Ask a question and get a dashboard. Ask a business question in plain language. Sigma finds the relevant data and builds an interactive dashboard, styled to match your company's brand. 2. Refine it in the same conversation. Adjust chart types, colors, and layout by describing what you want changed, without leaving the chat or starting over. 3. Add filters that make it useful. Add date ranges, segments, or other controls so the dashboard supports repeat analysis rather than a single snapshot. 4. Publish it, governed by default. When the dashboard is ready, publish it straight to Sigma, where it runs on live data with your existing account permissions applied. Why build the dashboard in Sigma? Sigma runs directly against your data platform: no extracts, no cached copies, no separate security model to keep in sync. When ChatGPT Work builds a dashboard through Sigma, it inherits the same governance, the same metric definitions, and the same live data as anything built inside Sigma itself. You can share and authenticate access to the dashboard while retaining peace of mind that each user only sees the data they're allowed to see. That's what keeps "everyone can build a dashboard" from turning into another version of shadow IT. Start building dashboards from ChatGPT Work. Existing Sigma customers have two ways to activate the workflow, depending on where your team works: * In ChatGPT Work, Sigma is included in OpenAI's Data agent. If a workspace admin has already enabled the Data agent, Sigma is there. * In ChatGPT or Codex, search for Sigma in ChatGPT's plugin directory and connect. The Sigma plugin is also listed through the Data agent in the ChatGPT Marketplace. Either way, sign in with your existing Sigma account. An Act, Analyze, or Build license and the permissions attached to it carry over, the same way they do with the Sigma MCP server. The dashboard building capabilities are available to Sigma users with the Build license, while the ask and query capabilities are available to all Sigma users. Learn more about using the Sigma plugin for ChatGPT in the Sigma documentation. New to Sigma? Request a demo to see the plugin build a dashboard on your own data. Frequently asked questions. What's different from the Sigma plugin I've been using in ChatGPT? Dashboard building. Search, exploration, and question-answering already worked. The update adds the ability to go from a question to a full interactive dashboard, then refine, filter, and publish it. Do I need a specific ChatGPT plan? Yes. This capability ships as part of OpenAI's Data agent update in ChatGPT Work, which requires a ChatGPT business, enterprise, or education workspace. Does my data leave Sigma's governance boundary? No. Permissions inherit from your existing Sigma account, so nothing bypasses the access controls already set on your connections, data models, and workbooks. What data can the plugin use? Anything you already have access to in Sigma: connections, data models, and workbooks where you have at least "Can view" or "Can use" permission. Can I keep editing the dashboard in Sigma after I publish it from ChatGPT? Yes. A dashboard published from ChatGPT is a full Sigma workbook. Open it in Sigma any time to keep building.
CloudEQS joins the Claude Partner Network: what it means for enterprise AI. * home * blog * CloudEQS joins the Claude Partner Network: what it means for enterprise AI. Overview. CloudEQS has formally joined Anthropic's Claude Partner Network. For most firms, that would be the starting gun. For CloudEQS, it's a formal recognition of work already in production - Claude has been part of how CloudEQS build Enterprise Intelligence Platforms for its customers well before this program existed. What the Claude Partner Network is. Anthropic launched the Claude Partner Network in March 2026 to support the consulting firms, system integrators, and AI specialists helping enterprises put Claude to work in production. Anthropic backed the launch with a $100 million investment covering partner training and certifications, technical enablement programs, joint go-to-market opportunities, and implementation frameworks for enterprise AI. Why now. CloudEQS's practice is built around Enterprise Intelligence Platforms on top of Snowflake and Databricks - and Claude isn't a new addition to that stack, it's already embedded across it. Snowflake Cortex is Snowflake's built-in suite of AI functions - including access to large language models like Claude - that lets teams run tasks like summarization, sentiment analysis, translation, and natural-language querying directly on data inside Snowflake, without moving it to a separate AI platform. Matillion MAIA is Matillion's AI-powered data pipeline agent that lets teams build, troubleshoot, and optimize ETL/ELT pipelines using natural language rather than manually configuring each step. Its engineers use Claude Code to write dbt scripts. Turning plain-language transformation requirements directly into tested, production-ready SQL models, cutting the manual effort of hand-coding and debugging dbt pipelines. Databricks Genie gives business users a way to ask questions about their data in plain English, and its partnership with Anthropic means those answers can be powered by Claude. Sigma Computing's Ask Sigma feature lets business users query live data with natural language instead of building dashboards from scratch, and its integration with Claude - through Sigma's MCP server and Claude connector - means those conversational insights are backed by the same enterprise AI reasoning teams already trust elsewhere in their stack. Formalizing its relationship with Anthropic doesn't start that work - it accelerates it. Structured training through the Partner Academy and a direct line to Anthropic's technical resources let CloudEQS up-level a team that's already fluent in Claude, build accelerators faster, and keep delivering ahead of where most of the market is still figuring out where to start. What's next. Moving forward, CloudEQS is focused on getting its team certified through Anthropic's Partner Academy and building Claude accelerators in CloudEQS Labs - reusable patterns for Snowflake Cortex, self-healing ETL pipelines, industry solutions, and migration converters that let CloudEQS move faster than a standing start would allow. Interested in enabling an Enterprise Intelligence Platform for your organization? Reach out to the CloudEQS team today.
Introducing App Templates: skip the blank page. Colin Dolese Product Manager July 8, 2026 Today Sigma Computing Inc. is launching App Templates in public beta. App Templates are a curated collection of Sigma-built templates covering common app use cases. Go from zero to a working app in minutes, connected to your own warehouse data. Starting from scratch is hard. Sigma gives teams everything they need to build powerful, AI-driven applications: Input Tables for writeback, Sigma Agents that take action, AI Columns to enrich and classify data, and action sequences that trigger external systems. But knowing how to structure those pieces for a new use case, or even knowing what's possible in the first place, isn't obvious from a blank workbook. Most teams end up doing one of two things: spending significant time figuring out patterns that others have already solved, or not starting at all. A revenue operations team building a territory management app has to hand-wire the input table logic, the layout, the controls, and the agent workflows from zero. A finance team tackling demand planning does the same. The patterns exist - but until now, there's been no good way to see them in a working context. That's what App Templates are for. Any time you're tackling a new use case, they give you a faster, better starting point than a blank canvas. How App Templates work in Sigma. Choose a template for a common use case. Sigma's template system lets you start from a working example rather than an empty workbook. App Templates are a new collection of Sigma-built templates covering common use cases, starting with 10 today: Each template opens with sample data pre-loaded so you can see exactly how the app works before connecting it to your own warehouse. Connect your own data. Once you've selected a template, connecting it to your own warehouse is a guided flow. Most of the sample data in each template is written directly to your warehouse using Sigma's writeback functionality - so swapping to your own data is largely a matter of pointing the connection to your own tables. A small number of common datasets, like users, will need to be mapped to your own equivalents. Learn the patterns that make apps powerful. Each App Template is a working example of Sigma's most powerful capabilities: Input Tables, agent workflows, AI Columns, and action sequences. Seeing these patterns in a working app is a fundamentally different experience than reading about them. For teams new to building in Sigma, App Templates are the fastest path to understanding what's possible. For experienced builders, they're reusable patterns - open a template, study the structure, and take what you need for your next build. Some templates are close enough to production-ready that teams can open one, connect to their own data, and ship with minimal customization. How Sigma Labs builds App Templates. Every App Template is built by Sigma's Labs team, a group of product experts who spend their time building real apps that can easily be adapted to meet each customer's unique needs. Each template is carefully crafted to reflect how a production app should actually be structured, and ships with a built-in ReadMe page and short video walkthrough so you can get oriented before you start customizing. The Project Management template, called Throughline, is a good example of what that means in practice. Throughline is a complete project management workspace, built entirely on Sigma Input Tables. It includes milestone planning, a Kanban board, ticket tracking, and a portfolio view across all active projects. A project manager can plan milestones, track execution, flag blockers, and communicate status to stakeholders, all from a single app where every page stays in sync. Update a ticket on the Board and it reflects immediately on the Home page and in the Plan. The Kanban board is one particularly strong example of the power of Sigma's app-building capabilities. It's a live, interactive board with columns for To Do, Blocked, In Progress, and Done. It's fully functional, built on Sigma's repeater element and Actions framework. Creating a new ticket triggers a multi-step flow that uses Sigma's AI capabilities to check for duplicate tickets before writing the new record to your warehouse. That's the pattern App Templates are designed to surface: how to put the pieces together in a way that actually works. Governed by default, warehouse-native throughout. App Templates are built on the same warehouse-native architecture as everything else in Sigma. When you open a template, Input Tables are created directly in your warehouse, so you're not working with copies or static files. Access controls, row-level security, and column masking you've already configured on your connection carry over automatically. There's nothing to re-govern when you go from template to production app. Start building with App Templates. App Templates are available in public beta for all Sigma customers today. Open a new workbook and select a template from the gallery to get started, or reach out to your account team to learn more. New to Sigma? Request a demo or start a free trial to see App Templates alongside the rest of the platform. Frequently asked questions. Can I customize an App Template after opening it? Yes. Every template opens as a fully editable workbook. You can modify any element: add columns, adjust the layout, change the agent logic, or extend the app with new capabilities. Can I still create my own templates? Yes. App Templates are a curated set of Sigma-built templates, but Sigma's template system supports custom templates too. Any workbook can be saved as a template and shared within your org. How is this different from Sigma Public? Sigma Public is a public sandbox where anyone can explore Sigma-built demos and community examples without logging in. App Templates live inside your own Sigma org - you connect them to your own warehouse data and own the output. What data platforms do App Templates work with? App Templates work with any warehouse connection in Sigma that supports writeback, including Snowflake, Databricks, BigQuery, and Redshift. See the documentation for details on source swap behavior by connection type. Are there any requirements to use App Templates? To take advantage of all AI app features included in App Templates, enabling Sigma's writeback and Sigma Agents functionality is recommended. But all App Templates can still be viewed and explored in all Sigma orgs even without these features.
Sigma Agents: introducing the new Agent Builder Assistant. Riley Gamboa Sr. Strategic Marketing Manager July 6, 2026 Building a custom AI agent in Sigma used to mean clicking through configuration screens and setting up each tool by hand. Now you can describe the workflow you want in plain language and let Sigma handle the setup. The new Agent Builder Assistant ("ABA" for short) is a natural language interface for designing Sigma Agents, the custom AI agents that run inside the security and governance of your cloud data warehouse. The Agent Builder Assistant is the first update in "Month of Agents," a July series where Sigma Computing Inc. release a new video each week covering a recent enhancement to Sigma Agents. Use ABA to: * Add a data source and get suggestions for agents you can build from its metadata * Describe a workflow in plain language and turn it into supported action tools * Generate and update your agent's instructions automatically Watch the full video above to see the Agent Builder Assistant in action, or read the accompanying transcript to understand how its Product, Design, and Engineering teams approached the ABA build. Of course, the best way to learn about Sigma Agents is to build one yourself. Sigma Agents are a public beta feature, and they're available by default to all Sigma customers that have enabled an AI provider. Reach out to your Sigma representative with any questions. If you're not a Sigma customer yet, you can request a demo or start a free trial to learn more. Next week in its Month of Agents series, Sigma Computing Inc.'ll cover governance, permissions, and using warehouse agents as tools. Return to the Sigma blog or follow its team on LinkedIn, X, or YouTube to catch the latest updates. Read on to see the full video transcript. What's new in Sigma's Agent Builder? Zalak: Hey, everyone. My name is Zalak Trivedi. I'm a Director of Product at Sigma. Over the next few weeks, Sigma Computing Inc. is going to be talking about agents: the way you build them, govern them, permission them, observe them, and a little bit about what's coming in the future. With me, I have Justine and Juwon, who are going to be talking about Agent Builder. You guys want to introduce yourselves? Juwon: I'm Juwon. I'm a Product Designer working on agents. Justine: I'm Justine. I'm a Software Engineer also working on agents. Zalak: Let's start simple. For someone who hasn't used Agent Builder in the last few weeks since the update, tell me what's new. Juwon: The new updates Sigma Computing Inc. has for agents give you a more intuitive way of building them. No matter who you are, whether you're a more technical user or an average business user, you can build an agent in a couple of clicks and a couple of prompts. Sigma Computing Inc. has also updated it so it's more data-source forward, so the first step you take is adding a data source to an agent so that agent has context. And Sigma Computing Inc. has the Agent Builder Assistant on the left (Sigma Computing Inc. call it ABA) to help you build the agents. How does the Agent Builder Assistant (ABA) turn natural language into agent tools? Zalak: In its first release, Sigma Computing Inc. had natural language instructions, which was pretty cool. I know people were ping-ponging between different systems for that. But Justine, you added natural language to tooling capabilities. Talk to me a little bit about that. Justine: Previously, you could set up your tools in the Agent Builder, but that required a few different clicks for manually configuring your tools. That process was a little tedious, because you'd be manually setting up every single configuration for each tool. What Sigma Computing Inc. introduced is that ABA is now able to build these tools for you. You write a couple of sentences in natural language, just describing your workflow, and ABA looks at that prompt, compares it to the list of supported tools Sigma Computing Inc. has, and translates your prompt into those tools. Zalak: Got it. Maybe give me an example. Justine: A common example is you want to filter your data down by some control value, and then you want your agent to run a query against that filtered data and return results. You describe this workflow to ABA, and ABA sees that Sigma Computing Inc. can set up a Set control value action tool, which updates your control value to whatever input the user requests. ABA also knows to update the instructions on the agent to describe how to actually use this tool, so the next time you pull up your agent, it knows exactly what to do. Zalak: Very cool. So it works with its entire actions framework. Justine: Exactly. Why did Sigma redesign the Agent Builder? Zalak: Juwon, before this redesign, I'm sure Sigma Computing Inc. thought through a lot of the user journeys. Where did people typically get stuck, and how are Sigma Computing Inc. helping them now? Juwon: One of the main pain points Sigma Computing Inc. wanted to solve with the Agent Builder redesign is that Sigma Computing Inc. wanted the agent-building experience to be as efficient and impactful as possible. Previously, users were kind of daunted coming up to Agent Builder, not necessarily knowing what the first step was or what agent they could make, and needing a couple rounds of prompting to get to the point they wanted to reach. With the new design, Sigma Computing Inc. is introducing a way to immediately get to an agent. For example, if you add some data sources using ABA, ABA can make an LLM tool call to suggest what agents you could possibly build. So within a couple of clicks, you're able to experience and interact with an agent, and then use that as a starting point for your agent-building experience. Zalak: Got it. So it's really for any technical level, from the marketing team all the way to somebody who knows how to write SQL. Very cool. Alright, Justine, do you want to show a quick demo of how this looks? Demo: How do you build a Sigma Agent from scratch? Justine: Let's do it. Hi, this is Justine, and Sigma Computing Inc. is just going to walk through what it looks like to set up your agent for the first time. I've got this headcount planning app here, which has some information about its open headcount requests. Let's say I want to add a new agent that helps users handle these requests, whether that's reviewing existing ones or creating new ones. Sigma Computing Inc. has got ABA on the left here, and its agent is blank, so it has some suggestions on how to get started. The first thing Sigma Computing Inc. want to do is add its data source. I'll go ahead and add the Input Table with all of its open recs in it, and Sigma Computing Inc. can see here that ABA was able to add that request table under its data sources. ABA can also look at the metadata associated with this table and suggest some types of agents it thinks Sigma Computing Inc.'d be interested in building. If Sigma Computing Inc. look at this list, this assistant option looks exactly like what Sigma Computing Inc. want. Once I select that, ABA goes ahead and builds its agent for Sigma Computing Inc., renaming it and filling out its instructions. You can see in the instructions it wrote that it's also able to directly tag that request table Sigma Computing Inc. just added. Its instructions are looking pretty good, but the other thing Sigma Computing Inc. want this agent to do is help with submitting new requests. Sigma Computing Inc.'ll just tell that to ABA in a couple of sentences: when the user asks to submit a new role, Sigma Computing Inc. should add that into its request table. ABA is going to think for a bit, but behind the scenes, it sees Sigma Computing Inc. is asking the agent to take an action. It reviews the list of supported action tools within Sigma and tries to match its request to one of those. It also fills out all of these configurations for Sigma Computing Inc. without Sigma Computing Inc. having to manually click into anything. It names its tool and writes a description. If Sigma Computing Inc. go back over to its instructions, Sigma Computing Inc. see it also knew to update its instructions to reference that new submit-request action tool it just created. In just a couple of prompts, Sigma Computing Inc. went from a completely blank agent to a fully functioning one, with access to data, tools to take action with, and the instructions to tie it all together. Zalak: Thanks, Justine, for a great demo. Next week in this series, Sigma Computing Inc. is going to talk about governance, permissions, and using warehouse agents as tools. Thank you, Juwon and Justine, for all the hard work on agents.