VectorMatch
Tech Area · AI & ML

AI & ML Engineering Recruitment

We recruit the engineers who take models and LLM applications from notebook to production: AI, Agentic AI, LLMOps, MLOps and Forward Deployed Engineers, and Data Scientists.

Talk to VectorMatch

How we assess each role

  • AI Engineer

    • An LLM-powered application they shipped: the model choice, prompts, and the product or API it sits behind
    • A RAG system they built: how they chunked and retrieved documents, and how they measured answer quality
    • Vector database work: embedding choice, indexing and similarity search tuned for recall and latency
    • Python
    • LangChain
    • Vector DBs
    • Embeddings
    • RAG
    • LLM APIs

  • Agentic AI Engineer

    • An agent or multi-step workflow they shipped: tool calling, state and memory, and where a human steps in
    • How they trace and evaluate agent runs, for example with Langfuse or OpenTelemetry, and catch loops or bad tool calls
    • Guardrails and cost or latency limits they put on agents in production
    • LangGraph
    • Tool calling
    • Multi-agent orchestration
    • Langfuse / OpenTelemetry
    • Evals
    • Python

  • LLMOps Engineer

    • An LLM feature shipped to real users, not a demo, and how they evaluated it before launch
    • Serving LLMs with Docker on Kubernetes: inference servers such as vLLM, with KV caching and batching to cut latency and GPU cost
    • The metrics they track: time to first token, tokens per second, p95 latency, cost per request, eval scores and hallucination rate
    • Dashboards and alerts they set up, for example in Grafana, and what triggers a rollback of a prompt or model change
    • Kubernetes
    • Docker
    • Grafana
    • vLLM
    • KV caching
    • Evals
    • Guardrails
    • Inference cost & latency

  • MLOps Engineer

    • Training and deployment pipelines they built or ran, containerised with Docker and running on Kubernetes
    • The production metrics they track: p95 latency, throughput, error rate and GPU utilisation
    • How they measure drift with KL divergence scores and PSI (Population Stability Index), and what triggers a retrain or rollback
    • Kubernetes
    • Docker
    • MLflow
    • CI/CD
    • Model serving
    • KL divergence & PSI

  • Forward Deployed Engineer

    • An AI integration delivered inside a customer's own systems and data
    • How they scoped the work with the customer and handled changing requirements
    • Whether they can explain a technical trade-off to a non-technical buyer
    • Full-stack delivery
    • APIs & integrations
    • LLM applications
    • Customer-facing scoping

  • Data Scientist

    • Regression and classification models they built and put to use, and how they chose features and evaluation metrics
    • Tree-based methods such as random forests and gradient boosting (XGBoost, LightGBM): when they use them and how they tune them
    • Deep learning work in PyTorch or TensorFlow, and how they validated models before handing them to engineering
    • Python
    • scikit-learn
    • XGBoost / LightGBM
    • PyTorch
    • TensorFlow
    • Statistics
Talk to us about an AI & ML Engineering hire

Open roles

All open roles

Hiring notes

  • Titles don't match skills yet

    "MLOps", "ML Platform" and "AI Engineer" mean different things at different companies. We scope the role by what the person will ship in the first six months, not by title.

  • Plan around notice periods

    Most senior engineers in India serve 60 to 90 days' notice. We share each candidate's notice period up front, so your start date is realistic.

  • LLM experience is easy to overstate

    Many CVs now list GenAI. We ask what the candidate put into production, how it was evaluated and what it cost to run.

Frequently asked questions

What is the difference between an MLOps and an LLMOps engineer?

MLOps engineers run the lifecycle of trained models: training pipelines, deployment and drift monitoring. LLMOps engineers focus on applications built on large language models: retrieval, evals, guardrails, and serving cost and latency. Many strong candidates have done both.

When do we need a Forward Deployed Engineer?

When your AI product has to be fitted into each customer's systems and data. An FDE sits between engineering and the customer and ships the integration.

Do you recruit for contract roles?

Yes. We recruit for full-time and contract roles.

Other Tech Areas

Ready to save hours of technical interview time on your next hire?

Partner with VectorMatch to receive vetted, interview-ready candidates tailored specifically to your technology stack and company standards.

Start Hiring Now