data & AI engineering
Data and AI Engineer — ML, Scorecards & Data Platforms
I build statistically grounded ML models and interpretable scorecards behind credit and collections decisions, and the AWS data pipelines that feed them.
Data and AI Engineer at AYE Finance, a regulated NBFC, and M.Sc. (Data Science & AI) candidate at BITS Pilani. 2nd Runner-up, AWS Agentic AI Hackathon; HackerRank SQL 5★.
01
XGBoost · ranking · pilot
Ranks overdue accounts by recovery likelihood, paired with an LLM-generated reporting layer that narrates segment trends. Live for a pilot group in real-user testing ahead of wider rollout.
02
WOE/IV · imbalance · monitoring
A Logistic Regression scorecard with WOE/IV binning, benchmarked against the XGBoost model. Class imbalance handled with SMOTE and cost-sensitive thresholds, which notably improved minority-class recall; PSI-based drift monitoring with MLflow versioning.
03
PySpark · AWS Glue · CDC
Design, build and own 8+ ETL/ELT pipelines (PySpark, Lambda, Glue) processing 4B+ rows across 5+ source systems, using CDC and a medallion architecture. Shipped the Data Warehouse 3 months ahead of deadline.
04
NL-to-query · LLM summaries
Added a natural-language-to-query layer to the Lending Analytics Platform (5,000+ employees, live production data) that turns plain-English questions into LLM-generated trend summaries.
05
QLoRA · benchmarking
Fine-tuned an open-source LLM with QLoRA for a lending-domain task and benchmarked it against prompted baselines on accuracy, latency and cost.
06
PostgreSQL · indexing · query plans
Tuned PostgreSQL queries behind the MIS dashboard through indexing and query-plan work, cutting report latency from 2 hours to 15 minutes.
Classification, clustering, WOE/IV scorecards, SHAP explainability, hypothesis testing, survival analysis, imbalanced classification (SMOTE), feature engineering
RAG pipelines, fine-tuning, LLM-generated reporting, natural-language-to-query analytics
Python, XGBoost, scikit-learn, PyTorch, Transformers, statsmodels, LangChain/LangGraph, AWS Bedrock, embeddings, Qdrant
SQL, Spark SQL, PostgreSQL, Power BI, query profiling and tuning, exploratory data analysis, dashboarding
PySpark, AWS Glue, CDC, medallion architecture, ETL/ELT design and ownership, Kafka, Airflow, data warehousing and modelling
FastAPI, REST APIs, Docker, Redis, AWS Lambda, Redshift, EMR, Step Functions, RDS
AYE Finance
Build the ML models and scorecards behind credit and collections decisions, and own the AWS data pipelines that feed them. Lead 6–7 engineers across data, ML and analytics initiatives through sprint planning, code review and stakeholder delivery.
AYE Finance
Led end-to-end design of a production data warehouse and ETL pipelines handling 10M+ records across LOS, LMS and CMS using SQL, PySpark and Python. Mapped complete data lineage, presented the architecture to the CTO and delivered an RBI compliance app 3× faster than planned with a 10-member team. Awarded “Cha Gaye — February” for delivery impact.
Accodyn Technologies
Built an ML-based risk detection model reaching 93% accuracy for a lending application, and contributed full-stack ML features and UI/UX.
RK Agrotech
Built an inventory and lead-management dashboard (HTML/CSS/JS + FastAPI) backed by PostgreSQL, with automated data ingestion from Google Forms and Sheets.
BITS Pilani (Digital)
In progress.
MIT-WPU, Pune
XGBoost · WOE/IV · SMOTE · MLflow · LLM
An XGBoost model that ranks overdue loan accounts by recovery likelihood, with an interpretable scorecard benchmark and LLM-written segment reports.
Python · scikit-learn · statsmodels · lifelines · Credit Risk
An interpretable WOE/IV credit scorecard benchmarked against gradient boosting, plus survival and causal analysis of time-to-default on a public lending dataset.
XGBoost · SHAP · FastAPI · ML
Prioritises loan applicants with an XGBoost model and explains each ranking with SHAP.
ETL · Python · DuckDB · Spark · YAML
A metadata-driven ETL framework that compiles YAML pipeline definitions into DuckDB, Polars or Spark jobs.
React · FastAPI · PostgreSQL · Analytics · RBAC
A live KPI dashboard for branch, zone and state operations at AYE Finance: React on the front, FastAPI and PostgreSQL behind it.
LangGraph · RAG · AWS Bedrock · Qdrant · FastAPI · Electron
Multi-agent RAG for credit and policy questions at AYE Finance, with guardrails, private document chat, and desktop and WhatsApp clients.