Job Description :

We are looking for an experienced Artificial Intelligence Engineer with strong hands-on expertise in building and deploying production-grade LLM and Agentic AI systems. The ideal candidate will have experience designing autonomous agents, orchestration frameworks, evaluation systems, structured-output workflows, and observable AI applications.

The role focuses on building reliable AI systems that can move from experimentation to shadow, assisted, and autonomous production environments, with strong emphasis on safety, evaluation, resilience, and operational efficiency.

Key Responsibilities :

- Design, develop, and own AI sub-agents with well-defined typed input/output schemas, versioned prompts, constrained toolsets, and measurable success criteria.

- Design and implement the agent orchestration layer, including planning, replanning, tool-failure recovery, state management, and handoffs between supervisors and sub-agents.

- Build evaluation and testing frameworks to assess agent quality, reliability, safety, and readiness for progression from shadow to assisted to autonomous operation.

- Develop mechanisms for structured outputs, schema validation, guardrails, and deterministic handling of model responses.

- Design multi-model architectures using high-capability models for complex reasoning, judgment, and classification, alongside lightweight/self-hosted models for high-volume extraction and formatting workloads.

- Build production-ready LLM applications with appropriate prompt engineering, model routing, tool integration, and failure handling.

- Implement observability for AI systems, including agent traces, execution metrics, failure attribution, latency, token usage, and cost per agent/workflow.

- Diagnose production failures and continuously improve agent behavior, system reliability, and model performance.

- Collaborate with Data, Software Engineering, Product, Security, and Platform teams to integrate AI capabilities into enterprise applications.

- Establish engineering standards for versioning, testing, deployment, monitoring, and lifecycle management of AI agents and LLM applications.

- Contribute to technical architecture decisions for scalable, secure, and maintainable AI platforms.

Required Skills & Experience :

- 6 - 8 years of experience in AI Engineering, Machine Learning Engineering, Software Engineering, or a closely related field.

- Proven experience building and deploying LLM-powered systems in production, beyond proof-of-concept or demo environments.

- Strong hands-on proficiency in Python.

- Practical experience with Agentic AI / agent orchestration, using frameworks such as LangGraph, workflow engines, custom runtimes, or equivalent approaches.

- Strong understanding of agent architecture, including supervisor/sub-agent patterns, tool calling, planning, state management, retries, handoffs, and recovery mechanisms.

- Experience building LLM evaluation (eval) frameworks, benchmarks, or quality-assessment pipelines.

- Strong understanding of structured generation, schema validation, prompt versioning, and reliable output handling.

- Experience implementing observability and monitoring for LLM/agent systems.

- Strong understanding of LLM application architecture, model selection, prompt engineering, and production reliability.

- Ability to diagnose and improve failures in AI systems using logs, traces, evaluation results, and production metrics.

- Strong software engineering fundamentals, including API design, testing, version control, CI/CD, and scalable application development.

- Strong analytical, debugging, problem-solving, and cross-functional communication skills.

Good to Have :

- Experience with Go or TypeScript.

- Experience with durable execution platforms such as Temporal, Cadence, AWS Step Functions, or similar.

- Experience with self-hosted inference using vLLM, TGI, quantization, GPU scheduling, or related technologies.

- Experience with large-scale retrieval systems, vector databases, and RAG architectures, including understanding when retrieval is appropriate and when it is not.

- Experience with document understanding, including OCR, layout analysis, document parsing, and extraction from complex PDFs.

- Experience building AI systems requiring audit trails, approval workflows, governance, or compliance reporting.

- Exposure to cloud platforms such as AWS, Azure, or GCP.

- Experience with AI security, guardrails, model governance, or responsible AI practices.