Senior Machine Learning Engineer

This is an AI platform engineering role wearing a machine learning title, and the distinction matters. The posting asks for 5+ years building large scale data, ML or platform systems with strong software engineering fundamentals in API design, concurrency and distributed systems. The work described is the infrastructure that production LLM systems actually need: hybrid retrieval combining graph, vector and keyword search with cache and TTL strategies balancing accuracy, latency and cost, vector stores such as pgvector, Qdrant or Milvus, and LLM gateway work covering provider adapters, routing policies, caching, quota and rate limiting, and token and cost accounting. The stated mindset is cost aware, treating latency and money as first class metrics and designing for graceful degradation. Apply if you build the platform under the models rather than training them. Bengaluru.

Required by the posting: 5+ years building large scale data, ML or platform systems, with strong software engineering fundamentals covering abstracted API design, concurrency and distributed systems. The requirement list also covers hybrid retrieval patterns and vector stores including pgvector, Qdrant and Milvus.

Highly relevant experience the posting calls out: having built or operated an LLM gateway or control plane, covering provider adapters, routing and policies, caching, quota and rate limiting, and cost and token accounting; powering retrieval with hybrid search across graph, vector and keyword, with cache and TTL strategy to balance accuracy, latency and cost.

The preferred mindset, which the posting sets out at unusual length and is worth reading as a description of the team's values: platform as product, meaning an obsession with developer experience, paved roads and clear SLAs; thinking in systems, where observability, fallback and access control are core rather than afterthoughts; genuine enthusiasm for enabling real world LLM and agentic use cases; cost awareness, treating latency and dollars as first class metrics and designing for graceful degradation; vendor agnostic thinking, choosing the right model or provider per use case and building for portability and resilience; and documentation and teaching, making complex systems understandable and upskilling teams.

Location: Bengaluru, Karnataka.

Benefits named on Tekion's postings: competitive compensation, generous stock options, and medical insurance coverage.

Honest fit guidance: do not be put off by "Machine Learning Engineer" if you have never trained a model. The requirement set is platform engineering, retrieval infrastructure and cost control, not research. Conversely, if you are a data scientist looking for modelling work, this is the wrong role, because nothing in it asks you to build models. Note that Tekion also runs a Staff Software Engineer AI Engineer requisition at 8+ years over a near identical brief, so the same work exists at a more senior band if you are past this one.