Principal AI Engineer (Python, LangGraph)

Quick Facts

• Remote (100%): Yes
• Engagement: B2B
• Rate: up to PLN 180/h
• Start: ASAP

Description

You’ll serve as the technical owner of a production, Python-based AI core for an LLM-driven mobile coaching platform. The system includes a FastAPI chat surface, a LangGraph multi-agent framework with orchestration, Celery + Redis async task flows, and OpenAI integration for the LLM layer. Your work centers on auditing the existing setup, defining the target architecture, and building evaluation/testing so changes don’t break coaching behavior or cause unpredictable token-cost growth.

Responsibilities

First 90 days

• Audit atlas-ai: agent flows, LangGraph state machines, Celery topology, datastore usage, OpenAI integration patterns
• Produce an operational risk assessment (failure modes, race conditions, retry semantics, idempotency, checkpoint integrity)
• Quantify token cost per agent flow and per user session
• Identify highest-risk subsystems and propose stabilization plans
• Build or harden an evaluation harness (golden cases, regression suites, hallucination/safety tests)
• Lead knowledge-transfer sessions from the client’s AI team

Ongoing

• Set the technical direction for the AI core
• Lead design for new agent flows and major changes to existing ones
• Own production health of the AI surface (with platform/SRE support)
• Hire and mentor the AI squad (~10 engineers)
• Represent the AI core in cross-team architecture discussions

Requirements

• 7+ years of Python in production at senior+ level
• Deep LangGraph experience (state graphs, checkpoints, interrupts, multi-agent supervision, subgraphs)
• Strong LangChain ecosystem knowledge (chains, tools, memory, output parsers, callbacks)
• Production FastAPI experience (streaming responses, dependency injection, middleware, async patterns)
• Production Celery + Redis experience (task ordering, retries, idempotency, priority queues, dead-letter handling)
• Strong Python concurrency skills (asyncio: gather, structured concurrency, cancellation; safe sync/async mixing)
• Multi-datastore operations with MongoDB + Redis + Postgres in one service, including transaction boundaries
• Experience scaling OpenAI API usage (rate limits, exponential backoff retries, fallback model routing, streaming, tool/function calling)
• Agent design patterns (ReAct, plan-and-execute, supervisor patterns, tool-use loops, multi-turn state, interrupt resumption)
• Prompt engineering discipline (evaluation, A/B testing, prompt version control, regression detection)
• Token cost optimization (prompt caching, model tiering, context trimming, summary memory)
• Production LLM observability (per-route token spend, prompt-level tracing, drift monitoring)
• Testing discipline (pytest, pytest-asyncio, property-based testing, snapshot tests for prompts, eval-based agent tests)
• Pydantic v2 fluency and type-hinted code throughout

Benefits

• 100% remote work
• B2B engagement with a rate up to PLN 180/h