Lead AI Engineer - Agent Framework & Agentic Applications
First seen by Alion on Sep 10, 2026.
Lead AI Engineer - Agent Framework & Agentic Applications
Build Ionixx's standardized agent framework - and the agents that run on it.
Role : Lead AI Engineer - Agent Framework & Agentic Applications
Location : Chennai - Work from Office
Type : Full-time
Level : Lead Engineer
Total Experience : 8 - 12 years
About the Role :
Build the Agent Framework :
- Architect and build the framework - the orchestration engine, a model-agnostic multi-LLM layer with routing and fallback, memory and state management, and a standard tool/integration layer (including MCP).
- Bake in the standard capabilities - guardrails and security, observability and evaluation, human-in-the-loop approvals, cost governance, and lifecycle/versioning - so every agent inherits them by default.
- Create reusable building blocks - templates, SDKs, and a clean developer experience so new agents are a thin layer of business logic, not a bespoke rebuild.
- Set the standards - patterns, guardrail policies, evaluation practices, and documentation that internal and client teams follow.
Develop Agents :
- Design, build, and ship production agents on the framework for real business use cases.
- Ground agents in knowledge (RAG) - retrieval over vector stores, context and prompt engineering, and integration with real systems (e.g., source control, ticketing, internal apps).
- Evaluate and continuously improve - build eval datasets, automated scoring (including LLM-as-judge), and regression testing; use production traces to drive quality up.
- Ship safely - versioning, staged/canary rollout, and rollback, with agents observable and cost-controlled in production.
Lead & Collaborate :
- Provide technical leadership - mentor engineers, review designs and code, and grow agent-development capability across the team.
What We're Looking For :
- Experience : 8 - 12 years in software engineering, with recent hands-on experience building and shipping LLM / agent applications to production.
- Machine learning : a solid grounding in ML fundamentals - model behavior, evaluation, and data. Enough to reason about LLM capabilities, limits, and fine-tuning trade-offs (applied ML/NLP or MLOps experience a strong plus).
- Programming : Strong Python (TypeScript/Node a plus); solid software-engineering fundamentals, system design, and API design.
- Agent frameworks : Hands-on with an agent-orchestration framework (e.g., LangGraph / LangChain, or equivalent) and LLM provider APIs (Anthropic Claude, OpenAI, and others), including open-weight models.
- Grounding & RAG : Retrieval-augmented generation, vector databases (e.g., pgvector, Pinecone, Chroma), embeddings, and prompt and context engineering.
- Observability & evals : Experience tracing and evaluating LLM apps (e.g., LangSmith, Langfuse, OpenTelemetry) and building evaluation/regression harnesses.
- Safety & security : Understanding of LLM-app risks - prompt injection, data leakage, least-privilege tool access - and how to guard against them.
- Platform : Cloud (AWS/GCP/Azure), containers, CI/CD, and building reliable, reusable services or SDKs.
- Leadership & communication : Ability to set technical direction, mentor, and communicate clearly with both engineers and business stakeholders.
Nice to Have :
- Experience with the Model Context Protocol (MCP) and building/integrating tool servers.
- Multi-agent orchestration, agent memory systems, and cost-optimization (model routing, prompt caching).
- Self-hosting or fine-tuning open-weight models; MLOps / LLMOps experience.
- Client-facing or consulting background; ability to shape and demo solutions for prospects.
- Domain exposure relevant to Ionixx's clients (e.g., fintech, healthcare) and open-source contributions.
Why Ionixx :
- Own a high-visibility, strategic initiative from the ground up - real scope and real impact.
- Work at the frontier of agentic AI, across internal products, demonstrations, and client delivery.
- Shape the standards and reusable IP that define how Ionixx builds agents.
Skills
Machine Learning, Python, LangChain, RAG, Prompt Engineering, Artificial Intelligence, Agentic AI, LLM