Looking for a job

Rahul T

@rahult

Generative AI Engineer at Moksha Infotech who developed legal RAG retrieval and drafting workflows with 630 passing tests.

India

Message

At Moksha Infotech, I contributed to a Legal AI system for Indian law, combining document processing, RAG-based retrieval, model workflows, and drafting support.

I built a retrieval workflow using query expansion, multi-pass recall, and ranking over statutes and precedents, with ChromaDB as the vector store. I also ran domain-adaptive pretraining of InLegalBERT on a curated Indian legal corpus.

For the Paperbook Compiler and Court Knowledge & Drafting Engine, I developed Python components for document processing and drafting, including PDF integrity checks and court-specific templates. I wrote pytest regression tests, and the suite reached 630 passing tests.

At Formonex, I built a vehicle-component failure prediction workflow using EDA, feature engineering, and ML models. I also automated vendor-data collection with Python and Selenium; earlier, at AiSPRY, I built a YOLOv8 hazard-detection model for industrial scrap images.

Experience

Work history, roles, and key accomplishments

Junior AI Developer

Moksha Infotech

Sep 2025 - Oct 2026 (1 year 1 month)

Contributed to a Legal AI system for Indian law, building retrieval workflows with query expansion and multi-pass recall, and ran domain-adaptive pretraining of InLegalBERT. Developed document processing and drafting components and wrote pytest regression tests, reaching 630 passing tests.

Python RAG ChromaDB InLegalBERT Hugging Face Transformers Query Expansion Multi Stage Retrieval Pytest Document Automation

Junior Data Scientist

Formonex

Jan 2025 - Sep 2025 (8 months)

Built a vehicle-component failure prediction workflow using EDA, feature engineering, and ML models to reduce operational downtime. Automated vendor-data collection from Google Maps with Python and Selenium across 13 categories.

Python SQL EDA Feature Engineering Machine Learning Selenium

Data Scientist Intern

AiSPRY

Jul 2024 - Aug 2024 (1 month)

Built a YOLOv8 hazard-detection model for industrial scrap images using transfer learning and Roboflow, and reviewed predictions with domain experts to reduce false positives.

YOLOv8 Transfer Learning OpenCV Roboflow