We are looking for a Senior AI/Machine Learning Engineer to set the technical direction for agentic AI on the Aera Platform. This is a role for someone who has already shipped autonomous agents into production, watched them fail in ways the demo never showed, and built the evaluation and guardrails that made them trustworthy the second time. You will make the architectural calls that the rest of the team builds on and you will be accountable for them when a recommendation moves a real supply chain.

This role will be based in our Pune office.

Responsibilities

- Set technical direction for agentic systems at Aera. Choose the architecture, write the design docs, and make the call on build-vs-buy, model selection, and framework adoption then defend those decisions with evidence when theyre challenged.
- Own the agentic platform end to end: multi-step reasoning, tool use, subagent orchestration, agent memory, and long-horizon autonomous loops including where a human checkpoint is non-negotiable and where it isnt.
- Own the evaluation discipline for the org, not just your own features. Offline and online evals, golden datasets, LLM-as-judge pipelines, regression gates tied to prompt and model versions, and test suites that hold up under non-determinism. Make it the default others inherit rather than the thing you did once.
- Establish the reusable substrate shared tool and skill surfaces, MCP servers, context strategies, prompt-caching and routing layers so common problems get solved once for the platform instead of repeatedly per team.
- Drive performance, cost, and reliability across the full set of levers: context engineering and compression, parallel and async tool calls, model selection and routing, structured outputs, token budgets, and inference optimization. Know which lever to pull from the trace, not from a guess.
- Raise the bar around you. Mentor engineers, review designs and agent-generated code with real rigour, and set the standard for what production-ready means for a non-deterministic system.
- Own agent security as a design property tool permissioning, sandboxing, prompt-injection resistance,



and the blast radius of an agent that misbehaves in a customer workplace.
- Partner with Data Science, Engineering, DevOps, and Product to turn ambiguous business problems into systems that hold up and push back when the framing is wrong.
- Evaluate emerging AI techniques and give the organization a defensible point of view on whats real, whats hype, and what we should adopt now.

About You

- B.E./B.Tech in Computer Science, Computer Engineering, or a related field.
- 5 8 years in software engineering and architecture, with at least 3 years designing and deploying ML or LLM-based systems and 12+ months building agentic or LLM-powered systems that reached production.
- You have set technical direction others followed. You can point to design docs you wrote, architectural decisions you owned, and at least one system you took from architecture through production reliability and then lived with including the incidents.
- You are a systems thinker. You look across business domains, find the problem thats actually being solved several times over, and abstract it into building blocks the whole platform can stand on. You step one click out from the problem in front of you and interrogate the assumption underneath it and you can bring other people to that altitude with you.
- You are a power user of agentic coding tools Claude Code, or equivalent agent harnesses with real intuition for where models are strong, where they fail, and how to tell the difference before it reaches production. You bring engineering discipline to agent-generated work: you review it, you gate it, you are accountable for it. We care that youve hit the failure modes, not that youve installed the CLI.
- You are fluent in current agentic engineering practice, not last years.



Context engineering, tool and skill design, subagent patterns, agent memory, evals and LLM-as-judge, structured outputs, prompt caching, RAG as one retrieval technique among several.
- You have opinions about evaluation that you can defend. You know why an agent that passes a benchmark can still be unshippable, and youve built the harness that caught it.
- Deep Python. FastAPI or equivalent for production services.
- Strong distributed-systems fundamentals; experience with large datasets and ML pipelines (Ray, Spark, or equivalent).
- Hands-on with PyTorch, Hugging Face, scikit-learn, pandas.
- Containerized microservices (Docker, Kubernetes) and CI/CD (Git, Jenkins, Jira).
- Humble and adaptable about code and frameworks. LangGraph or comparable orchestration frameworks are useful; none of them are the skill.
- You write and argue clearly. Senior work here is as much about bringing people to a decision as reaching it yourself.

Good to Have

- Experience leading a technical workstream or mentoring engineers, formally or otherwise.
- GoLang for high-performance components.
- Vector databases (Pinecone, Weaviate, FAISS, pgvector).
- Durable Execution platform like Temporal
- Event streaming and caching (Kafka, Pulsar, Redis).
- Agent observability and experiment tracking (Langfuse, LangSmith, OpenTelemetry, MLflow, WB, DVC).
- Fine-tuning where it genuinely beats prompting and context and the judgment to know when it doesnt.
- Multi-modal AI: text, image, and structured data in one workflow.
- Serverless AI infrastructure on AWS, GCP, or Azure.
- Enterprise-grade constraints: multi-tenancy, data isolation, compliance, or regulated deployments.

We need more systems thinkers, people who can look across all the business domains and abstract that to, heres the building blocks were going to need.

Disclaimer : This job posting has been aggregated from external source. Role details, content, and availability are subject to change. Applicants are advised to confirm the latest information directly on the company website before applying.

📌 Senior AI/ Machine Learning Engineer (Pune)
🏢 AERA SOFTWARE TECHNOLOGY
📍 Pune