Senior AI Product Engineer

About the role We build predictev™, an AI-driven audience intelligence platform used by pharma brand and marketing teams. The platform is live, the client list is growing, and the asks change weekly — a new brand, a new data cut, a new agent capability needed with a short turnaround.

We are looking for a senior engineer who can land inside an existing codebase and be productive in days. You will ship features on top of predictev™ under real client deadlines, make the call on what is worth building properly versus what needs to work by the deadline, and raise the technical bar for a small team of engineers and data scientists around you.

The differentiator we care most about: you have built for pharma commercial teams before, and you already understand what they are asking for.

What you’ll do

• Ship features on an existing platform — extend the predictev™ agent framework, data pipelines, and application layer without breaking what is already in production

• Turn client asks into working software fast — take an ambiguous request from a brand team and get a usable version in front of them in days, then iterate on real feedback

• Build and tune agentic AI features — retrieval, text-to-SQL, classification, and reasoning agents, with evals and observability rather than vibes

• Debug and stabilize — diagnose quality, latency, and accuracy issues across the stack and fix root causes, not symptoms

• Guide more junior engineers — code review, pairing, patterns, and unblocking, without needing formal authority to be listened to

• Pivot fast — priorities shift when a client or pitch shifts; you re-scope quickly and say clearly what moves out

• Make pragmatic architecture calls independently — balancing speed against technical debt, and flagging the tradeoff instead of hiding it

What we’re looking for

• 5+ years of professional software development experience

• Pharma or life sciences commercial experience — you have built analytics, data, or marketing products for pharma clients and understand how brand and medical teams work

• Proven speed on an existing codebase — you can read unfamiliar code, find the seam, and ship a change safely in the first week

• Hands-on AI-driven development — coding agents, LLM APIs, and AI-assisted workflows as a core part of how you build, not an experiment

• Comfort with client-driven ambiguity — requirements arrive incomplete and change mid-build; you ask two sharp questions and start

• Technical leadership instincts — you make other engineers better through review and example, and you escalate early rather than quietly slipping

• Strong written communication — remote, asynchronous, and reliable against milestones

• Strong judgment about when to optimize for speed versus robustness

Domain background we are prioritizing The strongest candidates have built for pharma commercial teams from inside a data, analytics, or consulting organization. Backgrounds we specifically want to see:

• Pharma data and analytics firms — IQVIA, ZS, Veeva, Komodo Health, Definitive Healthcare, Symphony, or similar

• Pharma marketing and media agencies — HCP and DTC campaign work, omnichannel planning, or brand strategy support

• Working fluency in the vocabulary — HCP segmentation and targeting, patient journey, NBRx/TRx, claims and script data, MMM and attribution, field and speaker program data

• Compliance instincts — PII and PHI handling, HIPAA and privacy constraints, and what makes a client security team comfortable

Tech stack

• Python — production services and data work; FastAPI or similar

• SQL and PostgreSQL — strong SQL is non-negotiable; large behavioral datasets, query performance, and schema design

• AI / LLM tooling — LLM APIs (Anthropic, OpenAI), agent frameworks (LangChain / LangGraph or similar), RAG and embeddings, prompt engineering, and an eval-first mindset

• LLM observability — Langfuse, LangSmith, or equivalent tracing and evaluation tooling in production

• Frontend — React and TypeScript, enough to build and fix the application layer end to end

• Cloud and infra — AWS (or GCP), Docker, and CI/CD; comfortable deploying your own work

• AI dev workflow — fluent with AI coding assistants and agentic dev tools (Claude Code, Cursor, Copilot, and similar)

Bonus points

• Background in fast-pa…