Looking for a job

Shawez Khan

@shawezkhan1

Message

Data Science Intern at Averixis Solutions who developed ML models reaching 88%+ accuracy.

India

Message

At Averixis Solutions Pvt. Ltd., I developed data preprocessing pipelines that improved dataset quality by 25% across 50,000+ records, and built and evaluated classification models achieving 88%+ accuracy on business tasks.

At Outlier AI & OneForma, I evaluated LLM outputs and contributed bias detection and quality benchmarks to training pipelines. I also built a PyTorch classifier for AI-generated faces that achieved 91.5% accuracy and a 0.93 ROC-AUC score.

Experience

Work history, roles, and key accomplishments

Current

AI Data Contributor - Linguist, Engineer & Analyst

Current

Data Annotation Tech

Aug 2026 - Present (2 months)

Complete code evaluation and software engineering annotation tasks across multiple programming languages, reviewing AI-generated code for correctness and logic. Perform rate-and-review assignments and general annotation tasks across NLP and domain-specific datasets.

Code Evaluation Ratings and Reviews NLP Linguistic Analysis Data Annotation QA

Current

Data Science Intern

Current

Averixis Solutions Pvt. Ltd.

Jun 2024 - Present (2 years 4 months)

Designed and executed end-to-end data preprocessing and cleaning pipelines in Python, improving dataset quality by 25% across 50,000+ records. Built and evaluated supervised ML models achieving 88%+ accuracy on real-world classification tasks.

Python Pandas NumPy Random Forest XGBoost Logistic Regression GCP BigQuery Vertex AI Matplotlib SeaBorn EDA

Current

AI Quality Engineer & LLM Evaluator

Current

Outlier AI & OneForma

Jan 2024 - Present (2 years 9 months)

Specialized in RLHF evaluation with 300+ verified hours, auditing AI-generated responses for factual accuracy, logical reasoning, and ethical alignment. Curated high-precision ground-truth datasets for computer vision and NLP classification.

RLHF LLM Evaluation Data Annotation Computer Vision NLP Quality Audits Ground Truth Validation