Open to opportunities

Prajwal User

@prajwaluser3

Message

I train, evaluate, and red-team LLMs using rigorous grounding, rating, and annotation workflows.

India

Message

What I'm looking for

I'm looking to contribute to AI training, model evaluation, prompt engineering, red-teaming, and RLHF workflows where rigorous evaluation data improves model quality.

I've trained, evaluated, and stress-tested Large Language Models at Telus International, designing multi-turn prompts and adversarial red-team tests to improve generative AI accuracy and contextual relevance.

I audit model outputs against grounding benchmarks, identify hallucinations and logical edge cases, and provide structured evaluation data to engineering teams. I've also documented evaluation metrics, data patterns, and system latency behavior.

Earlier at Telus International, I managed large-scale annotation workflows for autonomous-vehicle perception, social-media content evaluation, business-location verification, enterprise search intent, and AI training data.

I also build practical tools in Python, including a configurable LLM grading-rubric engine and a multi-agent research system using the Gemini API, async orchestration, live web-search grounding, and structured JSON prompting.

Experience

Work history, roles, and key accomplishments

AI Specialist

Telus International

Jan 2023 - Jan 2026 (3 years)

Designed and refined multi-turn prompts to improve generative AI accuracy, ran adversarial red-teaming to reduce hallucinations, and audited model outputs against grounding benchmarks. Authored technical documentation on evaluation metrics and system behavior.