At least 5 years of building production, data-heavy applications with Python, Django, and PostgreSQL.
Strong SQL skills beyond basic ORM queries, including aggregations, joins, and window functions, and the ability to design schemas and data models from use cases.
Experience designing and debugging asynchronous data pipelines, using Celery and Redis or equivalent tools.
Hands-on production experience with LLM APIs, prompt engineering, structured output, and cost management.
Fluency with agentic coding tools such as Claude Code and Cowork, and ability to use them to develop full-stack features.
Experience integrating third-party APIs with rate limits, authentication flows, and inconsistent data.
Ability to own frontend work using Django templates, HTMX, and Tailwind with AI assistance; frontend specialization is not required.
Strong written communication and ability to take ownership on a small, asynchronous team.
Responsibilities
Design data pipelines to pull, store, and structure performance, sales, and engagement data across hundreds of client accounts, including data from performance creative, TikTok Shop, TikTok LIVE, and Meta, while handling platform API limitations and inconsistencies.
Build an LLM analysis layer using Gemini for video analysis and tagging and Claude for generation, with structured output at scale and cost controls.
Build semantic search and discovery with pgvector so strategists can query creative insights.
Connect viral organic references, top-performing ads, content tags, and creator history into queryable creative intelligence accessible through the web app, Slack, and Cowork.
Develop creator-performance matching by linking creator history, content tags, and ad metrics to improve casting decisions.
Build and own dashboards, tag browsers, enrichment queues, and export tools using Django, HTMX, and Tailwind.
Integrate external APIs, including TikTok, Meta, Motion, and Cruva, and internal asynchronous pipelines using Celery and Redis.
Own features end to end, from ideation and data modeling through deployed user interfaces.
Desired Qualifications
Experience building agent harnesses, retrieval-augmented generation (RAG), or LLM evaluation pipelines, including side projects.
Background in or curiosity about machine-learning fundamentals, including when a simple scoring function is preferable to a complex model.
Experience with video or audio processing pipelines.
Domain experience in ad tech, the creator economy, performance marketing, or marketplace platforms.
Experience working with a US-based team in an asynchronous, English-fluent environment.
🏔️ About Sierra Studio
Sierra Studio is an innovative product development firm dedicated to empowering founders and business leaders in their quest to create impactful, meaningful products. Our team of experienced professionals is committed to delivering high-quality solutions that drive business growth. We prioritize a culture of good vibes, open communication, and continuous learning, providing our team members with the opportunities and resources they need to thrive.
This role is with one of the teams we help build through our staffing partnerships. When you apply, Sierra manages the hiring process, but you'll be hired directly by and work within the client company.
🏢 About The Company
Our hiring partner is a performance creative agency that makes ads for brands on TikTok, Meta, and TikTok Shop. They work with 2,000+ creators and produce thousands of videos per year. Their internal portal handles the entire workflow: creator onboarding, contracts, campaign management, content review, and payments.
They're now building a Creative Engine that turns creative strategy from gut instinct into a data-driven process. Data today lives in too many disconnected places — the Creative Engine brings it all into one central home, across performance creative, TikTok Shop, and TikTok LIVE, so the team can derive real insights instead of digging through separate platforms. AI watches and tags every video they produce, breaking each one into the hooks, formats, and patterns that drive results, and connects that to the sales and engagement data flowing out of Shop and Live. It all feeds one queryable system, a "brain" a strategist can ask "what are the top hooks for beauty this week?" and get a real answer back. The team reaches it through dashboards, Slack, and Cowork. It starts as the internal engine the creative team runs on, and over time clients log in to it too — this is the core of where the company is headed.
How they work:
Small team, high ownership. You own your work end to end, from idea (or data model) to deployed UI. No committees.
AI-assisted everything. Claude Code and coding agents are used daily as a real part of the workflow.
Ship constantly. They deploy to production multiple times per day. Scope stays small and tight.
Remote, async. Communication happens via Slack and GitHub. No standups or status meetings for the sake of meetings.
🖼️ About The Role
Our hiring partner is looking for its founding engineer on the Creative Engine, and they're hiring data-first, not app-first. This is a data product at its core, and you'll own the foundation: the pipelines that pull and structure performance data across hundreds of accounts, the LLM workflows that analyze every video they run, and the search layer that turns all of it into answers strategists can query.
You'll own features end to end, including the UI, but you don't need to be a frontend specialist. The heart of this role is the data and intelligence layer — you'll lean on Claude Code and agentic tooling to ship the dashboards and interfaces on top. Get the foundation right and the product follows, so they want someone who's genuinely LLM-fluent and savvy with AI-native development, not just someone who's heard of it.
🔨 What You'll Do
Design the data foundation: pipelines that pull, store, and structure performance, sales, and engagement data across hundreds of client accounts (performance creative, TikTok Shop, TikTok LIVE, and Meta), handling the messiness of real platform APIs.
Build the LLM analysis layer, using Gemini to watch and tag every video and Claude for generation, with structured output at scale and sane cost controls.
Build search and discovery with semantic search via pgvector, so a strategist can ask "what are the top hooks for beauty this week?" and get a real answer, not a keyword dump.
Build the "brain": connect viral organic references, top-performing ads, content tags, and creator history into queryable creative intelligence, accessible through the web app, Slack, and Cowork.
Develop creator-performance matching, linking creator history, content tags, and ad metrics to sharpen casting decisions.
Own the full stack on top of your data work: dashboards, tag browsers, enrichment queues, and export tools in Django + HTMX + Tailwind, moving fast with Claude Code (no dedicated frontend hire required).
5+ years building production, data-heavy applications with Python/Django and PostgreSQL (senior / lead level, autonomous, low-process).
Strong SQL well beyond basic ORM queries (aggregations, joins, window functions) and comfort designing schemas and data models straight from use cases.
Experience designing and debugging async data pipelines (Celery + Redis or equivalent).
Genuinely LLM-fluent: hands-on with LLM APIs in production, prompt engineering, structured output, and cost management.
Fluent with agentic coding tools (Claude Code, Cowork) and excited to use them to move fast, including building full-stack on top of your own data work.
Comfortable owning frontend (Django templates + HTMX + Tailwind) with AI assistance — you ship the whole feature, you just don't need to be a frontend expert.
Strong written communication and high ownership on a small, async team.
Preferred:
An AI engineer at heart — you've built agent harnesses, RAG, or LLM eval pipelines (even side projects count).
Background in or curiosity about ML fundamentals, like knowing when a simple scoring function beats a complex model.
Experience with video or audio processing pipelines.
Domain experience in ad tech, creator economy, performance marketing, or marketplace platforms.
Have worked with a US-based team before — async, English-fluent, knows the drill.
Location: remote, based in Brazil, working US time zones.