Website:
iamwarpspeed.com
Job details:
Company: warpSpeed (UK)
Location: Fully remote, India (with a daily overlap with UK working hours)
Job type: Full-time
Experience: 6+ years in software engineering, including 2+ years building LLM-powered applications in production
Salary: Highly competitive
About warpSpeed:
warpSpeed is building the world’s most personalised AI assistant. Our mobile app brings a person’s email, calendar, notes, tasks and messaging together across all their accounts, Gmail, Outlook and more and puts an AI assistant on top that understands how they work and helps them stay on top of everything.
We’re not building another inbox or another to-do list. We’re building one place where the AI can see the whole picture and act on it: triaging email, scheduling, drafting, following up and turning conversations into work that gets done.
Who you'll be working for:
warpSpeed is a UK-based company founded by Martin Warner, a serial entrepreneur with a track record of building and scaling technology ventures. It’s a founder-led team where engineering is a first-class priority: your work is visible, valued and central to every release.
About this role:
This is a full-time remote role for candidates based in India. You'll work directly with our UK-based leadership, product, and marketing teams on a product used by a global audience. Compensation is competitive within the Indian market and will be determined by your experience and expertise.
We’re looking for a senior AI engineer to own the warpSpeed AI assistant and lead a small team of two to three engineers working on it. You’ll report to the CTO, but day to day you’ll run AI engineering yourself: deciding how the assistant should reason, retrieve and act, making architecture calls, reviewing code and prompts, and proving with evaluations that changes actually make it better. The assistant is a multi-agent system in production. A LangGraph supervisor agent breaks down each user request and routes it to specialised sub-agents for email, calendar, tasks, notes, messaging, projects, web search, the product knowledge base and more. Those agents retrieve the user’s own data from vector indexes, call back into our platform to take actions, and return answers with references and suggested next steps. Your job is to make that system more capable, more reliable, faster and cheaper.
This is a hands-on engineering role, not a research role. We care about shipping AI features real people rely on every day, and about the software engineering that makes that possible.
What you'll do:
Lead AI engineering
- Take product requirements and turn them into agent designs, prompts, tools and well-scoped work for the team.
- Set and uphold engineering standards for AI code: testing, evaluation, logging and prompt management.
- Review pull requests and prompt changes, and mentor and unblock the other engineers.
- Own delivery: estimate honestly, flag risks early and keep the CTO informed without needing to be chased.
Own the assistant
- Design, build and improve the supervisor and sub-agents: routing, tool design, planning, memory and multi-step actions.
- Improve retrieval over users’ email, calendar, tasks, notes and messages: chunking, embeddings, hybrid and filtered search, ranking and context construction.
- Own the evaluation process. Grow our regression evaluation suites, measure against a real noise floor and make sure prompt and model changes are shown to help before they ship.
- Manage latency and cost: model selection, caching, token budgets and rate limiting. Every token we use turns into credits a user pays for, so accurate usage accounting matters.
- Build personalisation: help the assistant learn how each user works and adapt to it.
- Make the assistant safe to trust with someone’s inbox. Handle untrusted content (such as prompt injection hidden in incoming email), protect personal data, and make sure actions are taken only when the user intends them.
- Own production quality: tracing, observability, error handling and incident follow-up for the AI service.
Work across the product
- Work closely with the backend team on the contract between our Node.js platform and the AI service: what data the assistant sees, which actions it can take and how results come back.
- Work with product and design to define what the assistant should do, and be honest about what current models can and can’t do reliably.
- Keep up with a fast-moving field and bring in new models, techniques and tools when they earn their place.
What you'll Bring:
- 6+ years of professional software engineering, including 2+ years building LLM-powered applications that real users depend on.
- Strong Python: async programming, type hints, packaging and testing (pytest). You write production code, not just notebooks.
- Hands-on experience with LangChain and LangGraph, or equivalent agent frameworks, building tool-calling and multi-agent systems.
- Solid retrieval-augmented generation (RAG) experience: embeddings, vector databases, chunking strategies and retrieval quality tuning.
- Experience designing and running LLM evaluations (datasets, LLM-as-judge, regression testing), and the discipline to trust measurements over intuition.
- Strong prompt engineering skills, with an understanding of how prompts, tools and context interact across different models.
- Experience building and operating API services (ideally FastAPI) on a cloud platform, with observability and CI/CD.
- An understanding of LLM security and privacy risks, including prompt injection and data leakage.
- Experience leading or mentoring other engineers, whether as a tech lead, a team lead or the senior engineer others rely on.
- The ability to work independently in a fully remote team: you take ownership, communicate clearly in writing and don’t wait to be told what to do next.
- A degree in Computer Science, Engineering or a related field, or equivalent experience.
Our AI Stack
You don’t need to have used every one of these, but it helps to know what you’ll be working with:
- Python with FastAPI and uv
- LangChain and LangGraph, with a supervisor agent and a dozen domain sub-agents
- Google Gemini (via Vertex AI) as the primary model, with OpenAI embeddings
- AWS OpenSearch for vector search across per-domain indexes
- LangSmith for tracing and evaluation
- Valkey (Redis) for agent checkpointing, and MCP adapters for tool integration
- A background enrichment worker that processes new user data as it arrives
- pytest, Black, pre-commit, Bugsnag and GitHub Actions
- AWS ECS Fargate, deployed with AWS CDK
- A Node.js / TypeScript backend that the AI service is called from and calls back into
Nice to Have:
- Experience with Gemini or Vertex AI, and working across more than one model provider.
- Experience with LangSmith (or Langfuse, Braintrust or similar) for tracing and evaluations.
- Experience with OpenSearch or Elasticsearch, including hybrid lexical and vector search.
- Experience building AI features over email, calendar or other personal productivity data.
- Experience with agent memory, long-running agents or scheduled automations.
- Enough TypeScript / Node.js to work comfortably in our backend.
- Daily use of AI coding tools (Claude Code, Cursor, Copilot or similar), and good judgement about when to trust them.
What we offer:
- A highly competitive salary.
- Fully remote work.
- Real ownership: you’ll lead the AI assistant at the core of our product.
- An engineering-led, collaborative environment with a founder who values technical excellence.
- A product in daily use by a growing global user base, at the cutting edge of applied AI.
Click on Apply to know more.
