Staff Product Engineer, Data & AI

Prokeep

Staff Product Engineer focused on constructing data and AI systems for Prokeep's order automation sector. Engaging with messy datasets to improve product features and ensuring reliable outputs.

Posted 5/6/2026full-timePortland • Oregon • 🇺🇸 United StatesLeadWebsite

Tech Stack

Tools & technologies

ElixirPostgresPythonReact

About the role

Key responsibilities & impact

Take ownership of messy, heterogeneous, customer-generated data and design the systems, structures, and pipelines that turn it into a reliable foundation for product features.
Build systems that produce verifiably correct, structured outputs from ambiguous inputs — including the eval frameworks and ground-truth datasets that prove they actually work.
Ship customer-facing features end-to-end on top of that foundation, working across backend, data, and (where needed) frontend.
Make pragmatic decisions about when to reach for an LLM, when to use embeddings or retrieval, when classical ML is the right call, and when the answer is just well-modeled data and deterministic logic.
Own the quality and reliability of intelligent systems in production: latency, cost, fallback logic, monitoring, and continuous evaluation against ground truth.
Help shape Prokeep's strategy for where intelligence should live in the product, what to prioritize, and what tradeoffs to make.
Partner closely with our Data Engineer to surface and design around data quality issues, and with Product to ensure the systems we build actually move the business.
Improve systems already in production while contributing to greenfield initiatives — and over time, help us move from generalist LLMs toward custom models trained on our own data as it accumulates.

Requirements

What you’ll need

6+ years building and shipping production systems, with depth in backend and data work. Full-stack range is a plus.
A track record of taking messy, real-world data — sparse, inconsistent, shaped by inputs you don't control — and turning it into something a system can reason over reliably. You know how to profile, align, normalize, and structure it, and you don't get overwhelmed by scale or chaos.
Experience building systems where correctness is verifiable and measurable — search relevance, entity resolution, record linkage, catalog or taxonomy systems, data integration, or similar problems where there's a right answer and you have to prove you got it.
Comfort applying LLMs and embeddings where they're the right tool, and equal comfort using simpler, more deterministic approaches when they aren't. You've shipped features that used AI in production and you can speak honestly about where it helped and where it didn't.
Strong fundamentals in APIs, system design, data flow, and failure handling.
Pragmatic judgment about scope: how to limit a 45,000-row catalog to the 5,000 rows that actually matter, how to handle the long tail of edge cases, how to know when something is good enough to ship.
A strong product and customer mindset — you think in workflows and outcomes, not just systems.
Willingness to work in Elixir (our primary backend); strong experience with Python expected. Familiarity with React and PostgreSQL is a plus.

Benefits

Comp & perks

Competitive Compensation: Reflecting your expertise and impact.
Equity Package: Your success is our success—share in the growth you’ll help create.
Comprehensive Benefits: Health, dental, vision, life, short & long-term disability, 401(k), and employee assistance program (EAP).
Flexible PTO: Recharge and refocus with the flexibility to manage your time with no preset limits
Continuous Growth: Yearly education stipend to support your professional development.

ATS Keywords

✓ Tailor your resume

Applicant Tracking System Keywords

Tip: use these terms in your resume and cover letter to boost ATS matches.

Hard Skills & Tools

backend developmentdata integrationsystem designAPIsdata flowfailure handlingprofilingnormalizationentity resolutionrecord linkage

Soft Skills

pragmatic judgmentproduct mindsetcustomer mindsetdecision makingcollaboration