Senior AI Engineer (LLM systems)
Build production LLM systems for our customers: RAG, agents, evaluation harnesses, guardrails, and the observability that makes them trustworthy.
- Location
- Remote (EMEA / APAC overlap)
- Seniority
- Senior · 6+ years
- Type
- full-time
- Posted
- 2 April 2025
What you’ll do.
- Design and ship retrieval-augmented systems, copilots, and agentic workflows.
- Build evaluation suites and golden sets that gate prompt and model changes in CI.
- Implement guardrails: PII redaction, prompt-injection defences, scope controls.
- Instrument and operate LLM systems with per-request tracing and cost controls.
Requirements.
- 6+ years of software engineering experience, including 2+ on LLM systems in production.
- Strong Python and TypeScript fluency.
- Hands-on experience with at least two of OpenAI, Anthropic, Bedrock, or open-weight models.
- Practical experience with RAG, embeddings, and vector stores (pgvector, Pinecone, Weaviate).
Nice to have
- Background in classical ML or NLP before the LLM era.
- Published or open-source contributions to LLM tooling.
What to expect.
The process is designed to show how you think and collaborate in work that resembles the role — no puzzle interviews, no unpaid production assignments.
01 · Intro
Role fit and working style
A practical conversation about the role, time-zone overlap, compensation, and the kind of client problems you want to own.
02 · Work sample discussion
Bring evidence of judgement
Walk us through a shipped system, design case study, incident, repository, or architecture decision. We focus on trade-offs, not trivia.
03 · Peer review
Collaborate with senior teammates
You meet the people who would review your work. Expect questions about failure modes, communication, quality bars, and maintainability.
04 · Offer and first ninety days
Clarity before commitment
We share the engagement shape, expectations, benefits, compensation, and what success looks like in your first quarter.
The standards you will help uphold.
MayaLogic roles are senior because they include judgement, communication, and ownership of how software behaves after it leaves your laptop.
Role-specific quality bar
AI and data work ships with evaluation, lineage, PII handling, cost controls, and observability as first-class requirements.
Review culture
Every meaningful change is reviewed for behavior, security, maintainability, performance, and whether the client can own it.
Production responsibility
We write runbooks, instrument critical paths, and treat supportability as part of the role rather than someone else’s clean-up.
Send us a note via the contact form.
Mention the role title and link to whatever best represents how you work — a repository, a write-up, a system you are proud of. Until our applicant tracking system lands in Phase 3, the contact form is the fastest path to a real conversation.
Not quite the right role?
We still want to hear from you.
Send a note with the link to whatever best represents how you work. We keep an interest list for senior builders we would love to work with.