AI Lead

Role Overview

This position is for an AI Lead to guide the AI engineering stream of a large-scale initiative focused on creating an enterprise-ready Agent Development Platform. The role carries end-to-end technical ownership for the framework decisions, evaluation approach, and AI safety controls that will shape the platform and the solutions built on it. It is a hands-on leadership position, combining architecture, code review, and direct contribution to the most complex implementation work.

Technical Leadership and Architecture

• Lead the architecture decision-making process for the AI workstream, including the selection and approval of the agent framework, orchestration pattern, memory and state handling, and model routing approach.
• Define the standard agent blueprint for production use, covering planner and executor patterns, tool usage, human-in-the-loop checkpoints, and failure or compensation handling.
• Own the evaluation discipline, including ground-truth data design, accuracy measurement, calibration of LLM-as-judge methods, and regression checks integrated into CI.
• Take responsibility for AI safety and security, including defense against prompt injection, tool authorization boundaries, egress restrictions, and adversarial testing before release.
• Oversee the technical integrity of the accelerator component, including what is delivered as-is, what is customized, licensing hygiene, upgrade planning, and alignment with the broader roadmap.

Core Engineering and Delivery

• Build the highest-risk components directly, including the core agent SDK, evaluation harness, injection defense layer, and the first end-to-end production agent.
• Set the model selection, routing, and fallback strategy across providers through the LLM gateway, while also managing token spend, cost limits, and runaway-loop safeguards.
• Lead the knowledge architecture approach for retrieval and knowledge graphs, including embedding pipelines, hybrid search, and secure multi-tenant data access.
• Define the LLMOps and MLOps standards for tracing, monitoring, drift detection, quality checks, and release gating.

Team Leadership and Capability Building

• Manage and grow a pod of AI engineers, with accountability for technical quality, delivery outcomes, and team standards.
• Design the operating model for rapid agent delivery, with reusable templates, decision frameworks, and review gates that shorten delivery cycles.
• Mentor senior engineers, turn best practices into reusable internal assets, and contribute to broader AI capability-building efforts.
• Own technical interviewing and final-round assessment for AI engineering hiring.

Experience and Qualifications

• Bring 12 or more years of experience in software or AI engineering, including at least 3 years of production work on LLM-based systems.
• Have delivered at least one real agentic platform or multi-agent system handling live business transactions.
• Be highly proficient in Python and comfortable working with modern agent frameworks such as LangGraph or Claude Agent SDK, or similar tools.
• Have practical experience with LangChain and LangSmith, and familiarity with Prodigy is considered a strong advantage.
• Demonstrate strong evaluation engineering skills, including golden datasets, field-level accuracy, LLM-judge calibration, and hallucination measurement.
• Bring hands-on experience with prompt-injection defenses, authorization controls for tool calls, sandboxed execution, and adversarial testing of LLM applications.
• Have authored and defended architecture decisions for systems using durable orchestration, identity and multi-tenancy, vector or graph retrieval, and LLM routing layers.
• Show production-level experience with tracing, cost observability, CI/CD gating, and tenant-level cost tracking.
• Have led teams of at least 4 engineers through ambiguous, milestone-driven delivery work and helped grow senior talent.
• Be confident representing the technical solution to client CTOs and architecture boards, including defending decisions under scrutiny.

Preferred Background

Exposure to commercial real estate workflows such as lease accounting or CAM reconciliation is helpful, as is experience in document-heavy regulated sectors like legal, insurance, or financial services. Experience delivering under SOC 2, GDPR, or accessibility requirements, with FedRAMP exposure as an added plus, will be valued. Background in productizing AI platforms for resale or white-label use is also beneficial, along with public visibility through publications, open-source work, or conference speaking in the agentic AI space.

Additional Information

The role is based in Pune, India and is intended for a candidate who can take ownership of a high-stakes AI engineering track. Graduation is not required for eligibility.