Open to opportunities

Aditya Mehta

@adityamehta2

Chief Data Scientist leveraging Generative AI, credit risk modeling, and computer vision to drive measurable business growth.

India

Message

What I'm looking for

I’m looking to lead end-to-end AI/ML and Generative AI initiatives—building models, guardrails, and deployments that measurably improve risk, automation, and customer outcomes, with room to grow teams and technical standards.

I’m a Strategic Data Science Leader with over 11 years of expertise in AI and Machine Learning, specializing in Generative AI, credit risk modeling, and computer vision. I’m deeply hands-on with Python and PyTorch, building Generative AI workflows across prompt engineering, fine-tuning, RLHF, and agentic AI systems, then deploying SLMs and LLMs at scale.

In my recent roles, I’ve upgraded underwriting models to be aggregator agnostic, improving qualification and enabling major revenue growth, and I built a BERT-based enrichment engine to power purchase intelligence over 4B transactions using Milvus. I also developed an automated Stripe dispute/chargeback system using Claude Opus that cut Average Turnaround Time from 14 days to 1 day, and led teams to deliver OCR extraction and ML-driven automation that reduced manual effort and accelerated onboarding—while maintaining strong technical standards through training and boot camps.

Experience

Work history, roles, and key accomplishments

Current

Chief Data Scientist

Current

Line Financial Ltd

May 2025 - Present (1 year)

Upgraded the underwriting model to be aggregator-agnostic, increasing qualification by 40% and qualification amount by 20%, driving 200% revenue growth in 1 year. Built a BERT-based enrichment engine with Milvus for 4B transactions (95% categorization accuracy, 78% merchant identification) and deployed a Stripe dispute/chargeback automation using Claude Opus, reducing average turnaround time from

XGBoost Credit Risk Modeling BERT Embeddings Python

Chief Data Scientist

CapriGlobal Tech

Apr 2022 - May 2025 (3 years 1 month)

Led and scaled the Data Science team to eight members, running boot camps to onboard and upskill new hires. Developed AI credit risk models that reduced default rates by 2% and undercutting by 5%, and deployed an OCR-based extraction system with 95% accuracy that cut manual data entry costs by 40%.

Credit Risk Modeling Team Leadership BERT LLM Guardrails Python MLOps

Senior ML Engineer

One Assist Consumer Solution

Jan 2019 - Apr 2022 (3 years 3 months)

Automated visual damage and fraud checks for mobile insurance onboarding with a CNN pipeline, achieving 75% automation, saving 50 man-hours daily, and reducing ETA from 1–2 days to 5 minutes. Fine-tuned and deployed BERT for chatbot intent prediction, reducing fallback failures in Dialogflow by 65%.

Computer Vision Fraud Detection Dialogflow Python Redis Celery MLOps

Data Scientist

Jiva Ayurveda Ltd

Feb 2018 - Aug 2018 (6 months)

Developed and deployed a machine learning model to predict whether telephonic counseling orders require an audit, reducing audit requirements by 22%. Built a churn prediction model to identify users likely to discontinue consultation services.

predictive modeling Churn Modeling Feature Engineering Python Experimentation Data Processing

Assistant Manager (Data Science)

CitiCorp India Pvt Ltd

Jul 2015 - Feb 2018 (2 years 7 months)

Built an NLP solution to classify upstream log emails as critical, saving 40 man-hours monthly and reducing failure rate by 10%. Implemented an RPA automation framework (Automation Anywhere) to run processes in parallel, reducing license costs by up to $3,000 per month.

NLP Python Topic modeling scikit learn Data Processing (NumPy Pandas)