Principal AI Engineer (Python, LangGraph)

Quick Facts

• Role: Principal AI Engineer (Python, LangGraph)

• Engagement: B2B

• Location: 100% remote

• Start: ASAP

• Rate: Up to PLN 180/hour

Description

You will be the technical owner of a production AI-powered mobile coaching platform’s Python AI core. The product uses a FastAPI chat surface with tool-using agents, a LangGraph-based multi-agent orchestration framework, Celery + Redis for asynchronous flows, and MongoDB/Redis/Postgres for storage and checkpoints, with direct OpenAI integration. If the AI core breaks or costs spiral, product coaching quality and margins are impacted—so your work centers on reliability, evaluation, and cost control.

Responsibilities

• Audit atlas-ai: agent flows, LangGraph state machines, Celery topology, datastore usage, and OpenAI integration patterns
• Write an operational risk assessment covering failure modes, race conditions, retry semantics, idempotency, and 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 for agent flows (golden cases, regression suites, hallucination/safety tests)
• Lead knowledge-transfer sessions from the client’s AI team
• Ongoing ownership: set technical direction, lead designs for new/changed agent flows, and ensure production health of the AI surface with platform/SRE support
• Hire and mentor the AI squad (about 10 engineers at full scale)
• Represent the AI core in cross-team architecture conversations

Requirements

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

Benefits

• 100% remote
• B2B engagement
• Start ASAP
• Rate up to PLN 180/hour