G

Guild.ai

Multi-model AI agent governance platform

Forward Deployed Engineer

Full-Time
$180k - $300k/yr
Mid
San Francisco, CA, USA+1 moreMore locations: New York, NY, USA
HybridHybrid work; occasional customer-site travel of up to 25–30% may be required.
Software EngineeringAI Application Engineering, Full-Stack Engineering

About the job

Requirements
  • At least 3 years of technical pre-sales experience partnering with sales teams to position complex technical products and drive customer deals to closure.
  • At least 3 years of software engineering experience; full-stack development with Python and TypeScript is preferred but not required.
  • Deep comfort working with third-party APIs, authentication protocols, and enterprise systems.
  • Ability to navigate ambiguous requirements and translate complex business needs into clean, maintainable technical solutions.
  • Ability to explain complex infrastructure to non-technical stakeholders while maintaining technical credibility with developers.
  • Comfort shipping solutions in fast-paced, early-stage environments.
Responsibilities
  • Partner with customers to map business processes to agent-based solutions.
  • Design and build integration architectures connecting the platform to customer tools, including Slack, GitHub, Salesforce, and internal APIs.
  • Implement custom workflows and agent behaviors using modern large language model frameworks such as OpenAI and Anthropic.
  • Build supporting services, scripts, and interfaces to deliver customer solutions.
  • Work across the stack using Python, TypeScript, and React as needed.
  • Operate within cloud-native environments including Google Cloud Platform, Amazon Web Services, and Kubernetes.
  • Translate field learnings into concrete product requirements and platform features.
  • Identify patterns across customer engagements to build reusable templates and integration libraries.
  • Work with the core engineering team to connect customer needs with the platform roadmap.
  • Support sales with technical expertise, architecture reviews, and proof-of-concept builds.
  • Advise customer engineering and IT teams.
  • Travel occasionally to customer sites, up to 25–30%, to deepen partnerships and lead on-site implementations.
Desired Qualifications
  • Experience building with large language model APIs or agentic frameworks.
  • Background in developer tools, infrastructure, or enterprise software as a service.
  • Familiarity with Kubernetes and cloud-native scaling.

About the company

G

Guild.ai

View
Website

Guild.ai provides an infrastructure platform that helps organizations build, run, and manage many AI agents across different models and vendors. It offers a neutral control plane that unifies deployment, governance, and observability for AI agents, with features like centralized identity, access control, and complete audit logs. The platform includes components such as a Managed Agent Center and an Agent Hub for sharing trusted agents, plus tools for typed interfaces, versioned releases, and safe execution boundaries so agents behave like predictable systems rather than loose scripts. Originally starting as an open-source toolkit for automating machine learning experiments, Guild.ai now focuses on production AI governance and orchestration, aiming to create an ecosystem similar to GitHub for AI agents. The goal is to scale enterprise AI use by reducing shadow AI, improving reliability and collaboration across teams.

Company Size

11-50

Company Stage

Series A

Total Funding

$44M

Headquarters

Chicago, Illinois

Founded

2015

Get referred to Guild.ai

See people who can refer or advise you

Simplify Jobs

Simplify's Take

What believers are saying

  • June 2026 Insights Dashboard shows spend tracing, cache rates, and model-level cost control.
  • September 2026 public API and OpenClaw, LangChain support widen developer adoption.
  • March 2026 $44 million financing from GV and Khosla funds hiring and product expansion.

What critics are saying

  • OpenAI, Anthropic, and Google can bundle governance into their own platforms by 2027.
  • AWS, Microsoft, and Datadog already own enterprise control surfaces and distribution.
  • A weak agent ecosystem leaves Guild a feature, not infrastructure, killing standalone demand.

What makes Guild.ai unique

  • April 29, 2026 launch: code-first, model-agnostic control plane for AI agents.
  • Guild joins Linux Foundation’s Agentic AI Foundation, anchoring open-source governance credibility.
  • Agent Hub and Managed Agent Center package discovery, versioning, and sharing together.

Help us improve and share your feedback! Did you find this helpful?

Benefits

  • Health Insurance
  • Dental Insurance
  • Unlimited Paid Time Off

Growth & Insights and Company News

Headcount

6 month growth

0%

1 year growth

0%

2 year growth

0%
Guild.ai
Sep 20th, 2026
Guild Raises $44M to Build the Agent Control Plane

Guild AI raises $44M led by GV. CEO James Everingham on the agent control problem, what customers want, and Agent Hub.

Guild.ai
Sep 1st, 2026
Guild Optimizer: make your agents cheaper automatically.

Guild Optimizer: make your agents cheaper automatically. Ganesh Asapu Article Index AI agents are getting useful enough that the next problem is becoming obvious: how do you keep getting the same work done without continually spending more tokens to do it? It's easy to make an agent more capable by giving it a larger model, a bigger prompt, more context, and more tools. It's much harder to answer a different question. How much of that does the agent actually need? Today, Guild.ai, Inc. has added a new tool to Guild called Guild Optimizer. It's a new way to automatically reduce the cost of production agents while maintaining the quality of the work they produce. It's the next step in what Guild.ai, Inc. is building with Guild Insights, which helps you understand where AI resources are being consumed. Optimizer helps you act on it. From understanding spend to improving it. Most AI cost-management tools stop at visibility. They can tell you how many tokens you used, which models generated them, and how much they cost. That matters. But ultimately the useful question is 'Could we have achieved the same outcome for less?' Optimizer is designed to answer that question. Once an agent has been running on Guild and has accumulated just 10 production sessions, Guild can use that history to understand the work the agent is performing. From there, Optimizer can automatically test different ways of running that agent more efficiently. Instead of asking teams to manually rewrite prompts, switch models, remove tools, run tests, and compare results, Guild can do much of that work. The goal is simple. Production traffic becomes your tester. The difficult part of optimizing an agent isn't coming up with cheaper configurations. It's knowing whether those changes made the agent worse. That normally requires evals: a representative set of inputs and expected outcomes used to measure agent quality. Anyone who has built serious agentic systems knows creating and maintaining good evals can be a lot of work. Guild takes a different approach. Optimizer can use an agent's existing production sessions to automatically create an evaluation set representing the work the agent is already doing. That gives Guild a quality baseline. Guild.ai, Inc. can then change how the agent runs, replay representative work, and ask: Does the optimized version still perform as well as the current version? The initial evaluation can use an LLM-as-judge approach to score results, while giving teams the ability to inspect and adjust the evaluation criteria themselves. The important part is what happens next. Once you have an eval, optimization becomes measurable rather than speculative. Guild lowers costs in the following ways * Model Optimization * Tool Optimization * Prompt Optimization (coming soon) * Code Optimization (coming soon) Model Optimization: Try more efficient models - without guessing. One of the first things Optimizer tests is whether an agent needs the model it is currently using. A team might have selected a highly capable model when an agent was first built because it was the fastest way to get the workflow working. But that doesn't mean every production task requires it forever. Optimizer can run the same evaluation using a less expensive model and compare the results. At the same time, there are occasions when choosing a model with higher per token prices results in lower overall cost. A model with better reasoning can process the same workload with fewer tool calls. Fewer tool calls means fewer total tokens consumed, resulting in lower prices. Optimizer is designed to compare and contrast across those different dimensions, ensuring that if there are savings to be found, it will find them. Sometimes in unlikely places. If quality falls, nothing changes. If the lower priced model produces results that meet the existing quality bar, Guild can recommend the change and show the expected savings. This turns model selection from a one-time architectural decision into something that can be continuously tested against the actual work an agent performs. Tool Optimization: Give agents only the tools they need. Tools create another hidden source of agent cost. It's common to give an agent a large collection of tools so that it has everything it might need. But every additional tool adds context the model has to process, so removing unnecessary tools reduces the number of tokens used - even when the agent rarely or never uses it. Optimizer analyzes production sessions to determine which tools are actually being used. It can then test the agent with unnecessary tools removed. Again, the eval provides the guardrail. If the smaller toolset maintains quality, Guild recommends the more efficient configuration. Instead of asking teams to manually determine which capabilities an agent might safely lose, the system can test that question against real behavior. Prompt Optimization: Analyzes product behavior to propose a better replacement. Prompts tend to grow over time. Teams add instructions as new situations appear, but rarely revisit whether older guidance is still useful. The result can be a prompt that consumes unnecessary tokens, encourages unneeded tool calls, or leaves the model to spend extra turns deciding what to do. Prompt Optimization (coming soon) addresses this. It enables Optimizer to analyze the agent's current prompt alongside its production behavior and proposes a more cohesive replacement. That might mean removing instructions that no longer help, sharpening guidance that prevents unnecessary exploration, or combining steps that the agent currently performs separately. The goal isn't simply to shorten the prompt, but to reduce the cost of completing the work. Sometimes clearer instructions save more by preventing unnecessary steps. Guild then builds and evaluates the rewritten version against the current agent. Teams can inspect the before and after prompt, compare measured quality and cost, and decide whether to accept or reject the recommendation. Code Optimization: Use code when you don't need a model. Another important source of efficiency is determining which parts of an agent workflow need an LLM at all. Agents often begin life as large prompts because that is the fastest way to build them. Over time, however, repeated and predictable parts of those workflows can often be handled much more efficiently with deterministic code. Code Optimizer (coming soon) streamlines this part of the development process. Guild identifies opportunities to replace portions of an LLM-driven workflow with code while using the same evaluation framework to verify that the overall result still meets the required quality bar. That improves more than cost. Deterministic execution is faster and more predictable than repeatedly asking a model to perform work that software can reliably do itself. Optimizer isn't designed as a black box that silently changes production agents. For each optimization run, Guild produces a report showing the agents evaluated, the changes tested, and the impact on cost and quality. Teams can inspect recommendations individually and choose whether to accept or dismiss them. When a recommendation is accepted, Guild can install the optimized version into the workspace. And because Guild operates the agent runtime, this process can happen in the same system where the agents are already running. Costs reduced - in a single step. Guild.ai, Inc. has been using the same approach against its own agents at Guild. In one optimization run across a workspace containing four agents, Guild.ai, Inc. reduced the cost of a run from roughly $13 to $7.50 while maintaining the required quality. In another internal agent used for issue triage, optimization reduced the cost per run by approximately 87%, in part by moving work that didn't require an LLM into deterministic code. To be clear, those are individual examples rather than promises of a universal savings rate. Different agents have different workloads, models, prompts, tools, and opportunities for optimization. Substantial inefficiency can exist inside an agent that isn't visible simply by looking at its token bill. The only way to find it safely is to optimize against the outcome. Guild's different incentive model for AI infrastructure. There's also a broader reason Guild.ai, Inc. is building this. Model providers naturally benefit when applications consume more model inference. Guild's job is different. Guild.ai, Inc. want organizations to get the most useful work possible from the AI resources they pay for. That means helping you determine when a powerful model is necessary - and when it isn't. Which tools an agent actually needs - and which it doesn't. Which work requires intelligence from a model - and which work should simply be code. From agents you run to agents that improve. AI adoption will increasingly be measured by outcomes rather than token consumption. As companies deploy more agents, it won't be enough to know that they're running. Teams will need to understand what those agents cost, what they're accomplishing, and whether there's a better way to produce the same result. That's what Guild.ai, Inc. is building toward with Guild. Insights shows you what's happening. Optimizer helps you improve it.

Guild.ai
Aug 6th, 2026
Introducing Guild Agent Hub: one place to discover, run, and share production-ready AI agents.

Introducing Guild Agent Hub: one place to discover, run, and share production-ready AI agents. James Everingham Article Index The future of AI is not a collection of isolated assistants. It is a new generation of operational systems built around agents. Today, Guild.ai, Inc. is introducing Guild Agent Hub, an open platform for discovering, running, and sharing AI agents built for real work. There is already a growing ecosystem of open-source agents, skills, plugins, and automations being continually improved by developers around the world. Teams should not have to start from scratch every time they want to put one of those capabilities to work. Agent Hub gives developers one place to find useful agents, integrations, and skills, understand how they work, connect them to the right tools, and run them on the Guild control plane. Think of it as bringing together the openness of ecosystems with the production controls enterprises need to run them safely and reliably. An open platform for agents. The agent ecosystem is expanding quickly. Developers are already creating capable open-source agents for code review, automation, research, workflow management, and software development. Many companies also have repositories full of internal skills and lightweight agents that solve useful, repeatable problems. Agent Hub is designed to become the place where developers can: * Discover well-adopted open-source agents and skills * Install them without rebuilding the underlying infrastructure * Connect them to enterprise tools and data * Run them with the right identity, permissions, and controls * Share them publicly or across their organization * Fork and improve them for new use cases Guild is the way to discover open agents through Agent Hub and run them inside the enterprise. Over time, its goal is to make it possible to operate any useful agent through the same governed production layer. Start with what already works. Teams should not have to rebuild every agent from first principles. Agent Hub makes it easier to take those existing projects and turn them into capabilities that developers can run with one click on Guild. Find agents built for real workflows. Agent Hub is not a gallery of one-off prompt demos. It's a place to discover agents designed to solve practical problems across engineering, operations, productivity, and workflow automation. You can browse by use case, workflow, tool, or environment, then inspect an agent before deciding whether it fits what you are trying to do. The goal is simple: make it easier to start with something useful instead of rebuilding the same agent from scratch. Run agents on the Guild control plane. Finding an agent is only the beginning. To be useful in production, an agent needs to connect to real systems, operate with the right permissions, and run consistently across changing conditions. Guild provides the infrastructure required to operate those agents reliably, with visibility into what they are doing and control over how they interact with the rest of your stack. Share what works. The most useful agents should not remain trapped inside one project, repository, or team. Agent Hub makes it possible to publish agents publicly or share them internally, depending on how your organization works. Developers can fork agents for new use cases, improve existing implementations, and build on patterns that have already been tested by others. That means each successful agent can become a reusable building block rather than another isolated experiment. Move from experiments to production systems. The value of an agent is not measured by whether it can complete a demo. It's measured by whether it can become part of a dependable workflow. Agent Hub is built to surface agents that solve real problems, operate end-to-end, and can become part of the broader operational stack. The experience is straightforward: Browse. Find agents by use case, workflow, tool, or environment. Connect. Attach them to the tools, data, and systems they need. Deploy. Run them on the Guild control plane. Share. Publish agents for others to use, fork, and improve. Trusted and built for production realities. Production agents need more than intelligence. They need: * Visibility into what they are doing * Control over where and how they operate * Reliability across repeated workflows * Reusability so other developers and teams can build on what works Agent Hub brings those requirements together in one place. It gives developers a faster way to start with useful agents, and it gives organizations a clearer path from isolated experiments to governed, reusable systems. Discover agents. Run them on the Guild control plane. Share what you build with others.

Australian Associated Press
Jun 19th, 2026
Introducing the Guild Insights Dashboard.

Introducing the Guild Insights Dashboard. June 19, 2026 The first single view of how AI agents consume resources, drive usage, and incur costs SAN FRANCISCO, June 18, 2026 (GLOBE NEWSWIRE) - Tokenmaxxing or -minimizing might have dominated the conversation over the last few weeks, but at Guild.ai, the team knows the next phase of the token conversation is more about using tokens with purpose, visibility, and control as leaders ask the sharper question: what did all that spending actually deliver? Guild is building the control plane and infrastructure layer for the future, and today, they're adding to it with the introduction of the new Guild Insights Dashboard, which provides developers and organizations with detailed visibility into how their AI agents consume resources and incur costs. Organizations have watched their AI bills climb, and they've blown through their quarterly and even annual budgets, often without a full understanding of what exactly they've spent money on. Tech teams have responded by more closely scrutinizing every prompt, every model call, and every token spent. But focusing solely on token usage misses the bigger picture. Tokenmaxxing is simply an early signal of a much larger shift: organizations are beginning to operate hundreds of autonomous agents, yet most lack the tooling needed to observe, govern, and manage them reliably. The real problem isn't that organizations are consuming too many tokens. It's that they're deploying AI agents without the visibility and governance infrastructure required to manage them effectively. With Guild Insights, teams can now: * See spend in real dollars - exactly what AI is costing org-wide, not just raw token counts. * Trace it to the source - spend and tokens broken down by workspace, agent, user, provider, and model, each with its share of the total. * Track it over time - daily spend and usage across a 7/30/90-day window, compared to the previous period. * Spot what's avoidable - cache hit rate and input/output mix that show where cost is being wasted. The Insights Dashboard allows teams to figure out where they should focus their work on making their agents more efficient. By knowing how much they're spending on a particular agent, teams can determine what to improve to increase efficiency and impact. For teams that need deeper visibility, Guild Insights also provides agent-level breakdowns, allowing developers to understand exactly which agents are driving consumption. Organizations can further segment usage by provider, making it easy to compare spend across models and vendors. Whether a team is evaluating usage across Claude, Gemini, OpenAI, or other providers, leaders can quickly understand where resources are being allocated. This is only the beginning. Over time, Guild will continue expanding the visibility and governance capabilities available, such as: * Session-level usage analytics for deeper operational insights * Measure the token efficiency of agents * Budget enforcement at the agent and session level As organizations scale from a handful of agents to hundreds or thousands, the challenge will be managing an increasingly autonomous workforce of AI agents with the same level of visibility, accountability, and governance that companies expect from other critical business systems. With the Guild Insights Dashboard, organizations can confidently deploy AI agents, knowing they have the visibility and governance layers needed to manage them at scale. Learn more at Guild.ai. About Guild.ai Guild is the control plane for AI agents - a platform for building, deploying, governing, and sharing agents in production. It gives engineering teams a single place to manage how agents run, what they can access, and their costs. Guild is code-first, model-agnostic, and vendor-neutral, with security and governance built directly into the runtime. Teams can use pre-built agents, connect their own tools and APIs, and deploy custom agents using Guild's SDK. Guild has raised $44M, with a Series A led by Google Ventures, with participation from NFX, Acrew, Scribble Ventures, Khosla Ventures, and Webb Investment Network. Learn more at guild.ai. Media Contact: [email protected] AAPR aggregates press releases and media statements from around the world to assist its news partners with identifying and creating timely and relevant news. All of the press releases published on this website are third-party content and AAP was not involved in the creation of it. Read the full terms.

Yahoo Finance
Jun 18th, 2026
Guild.ai launches Insights Dashboard to track AI agent resource consumption and costs

Guild.ai has launched the Guild Insights Dashboard, providing organisations with detailed visibility into how AI agents consume resources and incur costs. The platform addresses the challenge of managing AI spending, allowing teams to track expenditure in real dollars across workspaces, agents, users, providers and models. The dashboard enables organisations to monitor daily spend over 7-, 30- and 90-day periods, identify avoidable costs through cache hit rates, and conduct agent-level analysis to improve efficiency. Teams can compare usage across providers including Claude, Gemini and OpenAI. Guild.ai, which has raised $44 million in funding led by Google Ventures, plans to expand capabilities with session-level analytics, token efficiency measurements and budget enforcement. The San Francisco-based company positions itself as a control plane for building, deploying and governing AI agents at scale.