Full Stack AI Engineer @ Data Science UA

- 5+ years of industry experience in product-centric full stack engineering roles;
- Very strong software engineering background with deep expertise in Python/Typescript.
- Hands-on LLM product patterns (prompting, tool use, routing, evals, regression testing);
Experience of working with Agents & LLMs:
- LangGraph/LangChain (know its limitations & trade-offs);
- Multi-provider routing, prompt tooling, trace/eval discipline (LangSmith);
Experience of working with Knowledge & Search Technologies:
- Vector — pgvector on PostgreSQL (semantic);
- Graph traversal — Cypher on Neo4j (relationships & enterprise context graph);
- Keyword/BM25 — lexical precision;
- Cross-Encoder (BGE Reranker) + RRF (blend lists robustly);
Deep knowledge of working with Data & Databases:
- PostgreSQL (Cloud SQL) — primary OLTP;
- SQLAlchemy and BigQuery — analytics/outcomes & cost/waste telemetry;
- Redis (Memorystore) — caching + lightweight queue/broker;
- GCS — object storage (files, audio, artifacts);
Production systems on GCP:
- Experience of working with Docker, Cloud Run (now) → GKE (later) and Pub/Sub;
APIs & integrations:
- Experience of working with FastAPI;
- Slack-native surfaces (human-in-the-loop);
- Enterprise adapters and bespoke integrations (existing MCPs and API platforms insufficient);
Security & Compliance:
- Experience of working with OAuth2/JWT;
- Least-privilege secrets (GCP Secret Manager);
Observability & Reliability:
- Structured logging, Cloud Logging; metrics/alerts → Cloud Monitoring;
- GitHub Actions CI/CD;
- Pre-commit/ruff/pytest, including Async.

Nice to have:
- Experience of working with Frontend: Next.js, Svelte, or SvelteKit;
- Experience of working with voice agents.
- Being observant in working with Datadog, Vertex/Bedrock rails for cost/perf resiliency.
- Previous experience of working in post-training (SFT/RL) or fine-tuning experience applied to real-world KPIs.
Deep knowledge in simulations & workflows:
- SimPy for discrete-event sims;
- Temporal or equivalent for long-running, reliable workflows.

,[ Tackle the end-to-end challenge of creating agents that solve open-ended enterprise problems., Wrangling high-volume enterprise data, building robust evaluation frameworks., Mastering the interplay between LLM prompting, fine-tuning, and scalable system architecture., Lead Customer Conversations & Translate to Product: Work directly with customer executives to translate their most critical objectives into breakthrough product features., Build the "New": Own the most audacious hypotheses, like fully autonomous optimization agents that replace entire consulting engagements and obsolete the need for army-style deployments., Master Deep Context: Build systems that understand enterprise architecture patterns better than most architects, automatically discovering waste and optimization opportunities., Take single-threaded ownership of core AI product areas, with end-to-end responsibility for customer success and business outcomes., Double as a Forward Deployed Engineer: Work regularly onsite with customers, and actually work with the systems., Every customer engagement becomes R&D for the platform, turning individual pain points into generalizable breakthroughs that serve the entire customer base] Requirements: Python, Full Stack, TypeScript, LLM, LangChain, LangGraph, PostgreSQL, BigQuery, Redis, GCS, GCP, Docker, FastAPI, CI/CD, VertexAI, Bedrock, SimPy, Next.js, Datadog

Kategorie

fullstack

👉 Mám zájem o práci