Job Description

Role Summary Build the “truth and control layer” for enterprise AI �� Why this role exists Most teams are busy building AI. We are focused on something harder: How do you evaluate, control, and trust non-deterministic AI systems at scale? At Hexaware’s AI Center of Excellence, we are building a research-driven applied AI unit, st-ABEL (Structural Theory for Agentic Behavior and Evaluation Lab); focused on: · Model evaluation and finetuning · Agentic systems and behavioral control · Evaluation-as-a-Service platforms We are looking for leaders who can think from first principles and shape AI thinking —not just use what exists. �� What you’ll work on · Design LLM finetuning and RL pipelines (SFT, DPO, RLHF/RLAIF) · Build evaluation frameworks for models, agents, and AI systems · Architect agentic systems (multi-agent workflows, tool use, planning) · Create simulation environments and synthetic evaluation setups · Define benchmarks and reliability standards for enterprise AI · Translate research ideas into production systems · Shape AI strategy (cost, performance, token economics, infra trade-offs) �� We are hiring for 2 complementary profiles: · Model Intelligence Lead → finetuning, RL, evaluation science · Agentic Systems Lead → agents, behavior, system design The Principal AI Research Consultant is responsible for designing next-generation AI systems with a focus on model evaluation, finetuning, agentic behavior, and system-level reliability. The role anchors the organization across model truth and system control and is expected to translate research concepts into production-grade systems, tools, and reusable platforms. This role requires a combination of deep technical expertise, systems thinking, and the ability to convert ideas into scalable tools and frameworks. · Aligned with st-ABEL’s identity as a research + applied systems lab Key Responsibilities