ML / AI Engineer (agents, RAG, production)

Who we need

An ML / AI Engineer at JimmyNeuron. You’ll work on conversational agents, RAG, answer-quality evaluation, and embedding models into products and client systems (including on-premise and closed networks).

We need someone who can take AI to production: latency, cost, guardrails, hallucination monitoring, human escalation - not just a pretty Colab prototype.

Responsibilities

• Design and ship LLM agents, RAG, and document/dialog processing pipelines
• Connect models to products: APIs, queues, cache, limits, prompt versions
• Build evaluation: quality metrics, regressions, A/B hypotheses on answers
• Cut inference cost and raise reliability (retry, fallback, handoff)
• Work with product backends and integrations (CRM, messengers, telephony)
• Join client pilots: vision, local LLMs, knowledge bases - by project

Who we’re looking for

• Hands-on experience with LLMs, embeddings, vector DBs, orchestration (LangChain / custom pipelines - either is fine)
• Python and/or TypeScript for production code, not notebooks only
• Understanding of model limits and ability to design around them
• Experience shipping AI to production in at least one commercial product / engagement
• Plus: speech-to-text / call NLP, CV, on-prem GPU, daily Cursor use

What we offer

• Work at the product + custom intersection: ChatNeuron, CommBoost, and client pipelines
• Access to real data and metrics (under NDA), not toy datasets
• Comp and level - based on experience and case portfolio
• Format: Kazan or remote, full-time