Open to opportunities

Aayush Jha

@aayushjha1

Message

AI data quality and evaluation specialist with cross-domain expertise in math, ML, and software engineering.

India

Message

What I'm looking for

I’m looking for a role in AI training, evaluation, or data-quality work where I can design and verify high-rigor problems, catch reasoning failures, and improve reward-model outputs using strong math, ML, and software engineering foundations.

I’m a pre-final year B.Stat. Data Science student at the Indian Statistical Institute, Kolkata (AIR 149, top ∼0.5% nationally) and I work on concurrent professional AI training contracts where my acceptance rate is consistently >95% across 100+ evaluated tasks. I qualify as a cross-domain expert to evaluate, design, and verify AI training content across Mathematics, Statistics, Computer Science, Machine Learning, Data Science, Software Engineering, and Financial Mathematics—always delivering rigorous, English-first work.

In my current role as an AI Data Quality Analyst (RLHF & Problem Design), I design original multi-step problems, verify AI-generated solutions for correctness and logical constraints, and apply structured scoring to catch edge-case hallucinations and discrete reasoning failures using Python (NumPy, SciPy, SymPy, scikit-learn). As a Writer for Model Response Evaluation (MRN), I run preference annotations with a 6-dimension rubric and provide clear written justifications for instruction following, truthfulness, correctness, writing quality, verbosity, and overall quality, strengthening reward model training. My applied work spans credit card fraud detection, a physics-informed Neural ODE system for hydrology with a major inference-speedup, and nonparametric survival analysis from scratch.

Experience

Work history, roles, and key accomplishments

Current

Model Response Evaluation Writer

Current

Mer•or

Jan 2026 - Present (5 months)

Evaluated LLM responses using a 6-dimension rubric (instruction following, truthfulness, correctness, writing quality, verbosity, overall quality) to generate preference annotations for reward model training. Performed rigorous fact-checking and logical soundness verification for multi-step math, statistics, and software engineering claims with written justifications.

LLM Evaluation Truthfulness Checking Technical Writing Preference Annotation

Current

AI Data Quality Analyst

Current

Ube• AI Solutions

Nov 2025 - Present (7 months)

Designed original multi-step RLHF training problems across math, CS, data science, software engineering, and financial mathematics, creating tasks that require non-trivial reasoning chains. Verified AI solutions for correctness and logical constraints using Python and maintained a >95% acceptance rate across 100+ reviewed tasks.

Problem Design Python NumPy Scipy scikit learn

Education

Degrees, certifications, and relevant coursework

Indian Statistical Institute, Kolkata

Bachelor of Statistical Data Science, Statistical Data Science

Activities and societies: Coursework includes Real Analysis, Linear Algebra, Probability Theory, Statistical Inference, Stochastic Processes, ML, Convex Optimisation, Combinatorics, Numerical Analysis, Graph Theory, and Database Systems. Projects include cost-aware credit card fraud detection optimization, Neural ODE FloodRisk system, and nonparametric survival analysis (Kaplan-Meier/Nelson-Aalen) with variance derivations.

Pre-final year student in the Bachelor of Statistical Data Science program at the Indian Statistical Institute, Kolkata (AIR 149, top ~0.5% nationally). Coursework covers core mathematics, statistics, machine learning, optimization, databases, and related topics.