Craft bespoke software solutions by designing, coding, and enhancing system components. Leverage modern frameworks and agile methodologies to deliver scalable, high-performance applications precisely aligned with unique business needs. This role emphasizes innovation and continuous improvement in software product development.

Collaborate effectively with team members, adapt to evolving project requirements, and contribute to a dynamic workflow. The position is situated in our Bengaluru office, requiring a commitment to on-site presence.

- Design and implement sophisticated multi-agent orchestration pipelines, utilizing LangGraph state machines for conditional routing, self-healing retry mechanisms, and supervisor/sub-agent hierarchies. - Engineer robust retrieval-augmented generation (RAG) systems, encompassing chunking, embedding pipelines, vector search, re-ranking, and efficient token-budget management. - Model knowledge graphs, defining entities, relationships, and ontologies, and construct hybrid retrieval systems that seamlessly fuse graph traversal with vector similarity. - Govern prompts through a versioned prompt registry, implementing multi-model routing based on task/stage with effective fallback strategies. - Build comprehensive LLM evaluation harnesses, including golden datasets, automated scoring, regression gates, and rigorous hallucination/grounding checks. - Implement advanced guardrails for schema-constrained (structured/JSON) output, validation, self-correction, and safety filtering. - Serve and optimize on-premise LLMs using vLLM, alongside cloud-based OpenAI-compatible APIs, managing batching, streaming, and context-window constraints. - Instrument the AI stack for comprehensive observability, tracking token usage, latency percentiles, cost attribution, and detailed trace/provenance logging. - Integrate agentic interoperability protocols (A2A, MCP) and connect agents to enterprise tools via REST/GraphQL interfaces.

- Possess a strong foundation in Python 3.10+, with expertise in production asynchronous code and FastAPI. - Demonstrate proficiency in LLM engineering, including prompt design, token budgeting, multi-model routing, structured output generation, and self-healing retry loops. - Showcase hands-on experience with agent frameworks such as LangGraph or LangChain, covering state machines, tool calling, and orchestration. - Exhibit practical knowledge of model serving technologies like vLLM, TGI, or OpenAI-compatible inference APIs, supporting streaming and batching. - Strong experience in LLM evaluation, including creating evaluation datasets, automated scoring, regression testing, and assessing hallucination/grounding metrics. - Proven ability to implement guardrails, encompassing output validation, schema enforcement, and safety filtering. - Familiarity with observability tools and techniques for tracking token usage, latency, costs, and system tracing. - A minimum of 7-8 years of professional software engineering experience is required, with at least 3 years dedicated to building LLM, agent, or ML systems in production environments. - Fluent communication skills are essential for effective collaboration and technical discussions.