Headline

AI Trainer & LLM Evaluator | RLHF · Code Evaluation · Red-Teaming | Python · C++ · JavaScript · SQL | 2,500+ hrs Evaluating Model Outputs | IIT Madras

About

I evaluate and improve large language models, with a focus on code and financial reasoning. Since May 2024 I've logged 2,500+ hours as an AI Evaluator at DataAnnotation. I qualified by assessment into two specialist streams, programming and financial analysis, and have kept my access through continuous quality scoring. My work includes ranking model outputs for correctness across C++, C#, Python and JavaScript, writing step-by-step reference solutions built for reviewer audit, and red-teaming models with adversarial prompts that expose logic errors, buggy code and flawed financial calculations. Before that, I spent 7 years as a freelance software developer building games in C++ and Unreal Engine. I programmed graphics, audio, UI, gameplay systems and custom engine plugins, and delivered against written client briefs on fixed deadlines. That's where I learned to read code critically. It's also why I can quickly spot where a model's solution will break. Background: B.Tech, IIT Madras; MA Game & Interactive Media Design, Brunel University London. What I bring: rigorous rubric-based evaluation, strong written reasoning, fast and consistent turnaround, and hands-on engineering experience. Open to: AI Trainer, LLM Evaluator, AI Data Specialist, Model Evaluation, and C++/Python Software Developer roles. Remote or Bengaluru. Core skills: LLM Evaluation · RLHF · SFT Data · Red-Teaming · Prompt Engineering · Python · C++ · C# · JavaScript · SQL · Financial Analysis

Location

Bengaluru

AI stack

langchain

langgraph

Work experience

AI Trainer & LLM Evaluator - Code & Financial Analysis at DataAnnotation, May 2024 - Present · 2 yrs 6 mos, Bengaluru, Karnataka, India, • 2,500+ hours over 28 months (20+ hrs/week) evaluating and improving LLM outputs for code and financial reasoning, alongside concurrent client work. • Qualified by assessment into two specialist streams (programming and financial analysis); access maintained through ongoing quality scoring. • Evaluated and ranked model responses for correctness across C++, C#, Python and JavaScript, producing preference data for RLHF. • Authored gold-standard reference solutions with step-by-step reasoning (SFT data), written for reviewer audit. • Red-teamed models with adversarial prompts targeting logic errors, buggy implementations and faulty financial calculations, scored against structured rubrics under fast turnaround. • Evaluated financial outputs covering accounting, valuation and quantitative reasoning. • Self-QA'd every submission against the rubric, re-checking constraints and edge cases before delivery.

Software Developer - Games (C++ / Unreal Engine) at Self-Employed, Sep 2017 - Sep 2024 · 7 yrs 1 mo, Bengaluru, Karnataka, India, Programmed graphics, audio, UI, gameplay systems and custom engine plugins across client projects. Worked to written client briefs across async revision cycles, delivering to agreed dates and incorporating structured feedback between milestones.

Education

Bachelor of Technology (B.Tech), Chemical Engineering, Indian Institute of Technology, Madras

Master of Arts - MA, Game and Interactive Media Design, Brunel University of London