AI Evaluation Lead (Part Time Contract, Engagement Based)

AI Evaluation Lead (Part Time Contract, Engagement Based) Dataclap Digital Ventures | Bangalore, Chennai or Coimbatore (remote first, occasional visits to Coimbatore) About us Dataclap is an AI data services company based in Coimbatore, working with AI companies across North America and Europe on data annotation, human in the loop workflows, RLHF and MLOps.

We're building a dedicated AI evaluation practice covering LLMs, AI agents, computer vision and audio models, and we're looking for experienced practitioners to lead client engagements and mentor our engineering team. The role This is a part time, engagement based contract role, not full time employment. You'll lead AI evaluation projects for our clients, typically 8 to 15 hours a week per engagement.

You'll design the evaluation approach, guide a team of junior engineers who do the hands on execution, review their work, and own the quality of what we deliver. You'll also help us turn every engagement into reusable methods, so the practice gets stronger with each project.

What you'll do

understand the model, the use case, and what "good" looks like for them Design evaluation frameworks, including metrics, test sets, rubrics, and pass/fail criteria Build or guide the building of golden datasets and evaluation harnesses Design and calibrate LLM as a judge setups, and know when human evaluation is needed instead Plan red teaming, safety and robustness testing where relevant Lead and mentor 2 to 5 junior engineers per engagement through regular reviews and working sessions Review results, catch flawed methodology, and present findings to client stakeholders Document each engagement as a playbook, rubric set or template the team can reuse What we're looking for 6+ years in machine learning, data science or applied AI, with at least 2 years focused on model evaluation, benchmarking or quality assurance of ML systems Hands on experience evaluating LLM applications: RAG systems, chatbots, agents or fine tuned models Solid grounding in evaluation fundamentals: metric selection, test set design, statistical significance, inter annotator agreement, and data leakage Experience with eval tooling such as Deep.

Eval, Ragas, Promptfoo, Inspect, Langfuse, Braintrust or EleutherAI's evaluation harness (or custom built equivalents) Experience leading or mentoring engineers, and comfort reviewing others' work critically Clear written and spoken communication; you'll present findings to international clients Python proficiency Strong plus Computer vision evaluation experience (detection, segmentation, tracking; metrics such as mAP and IoU; tools such as Fifty.

One or CVAT) Audio or speech model evaluation (word error rate, speaker diarization, MOS style human ratings) Experience with AI agent evaluation: task completion, tool use, multi step trajectories Familiarity with AI governance frameworks such as the EU AI Act or NIST AI RMF Prior consulting or client facing delivery experience Engagement model Contract basis, paid per engagement, with a monthly retainer during active projects Flexible hours, mostly remote, with occasional in person sessions in Coimbatore Start with a short paid pilot project so we can both see if it's a fit If you're currently employed, please make sure your employer permits outside consulting work How to apply Send us a short note (not just a CV) describing one model evaluation you designed or led: what you were evaluating, how you decided what to measure, and one thing that surprised you in the results.