Highly experienced Full-Stack Data Scientist / Machine Learning Engineer and Software Developer, delivering production ready ML models and customised softwares to automate business processes and decision making. With an added advantage of attaining MBA from India's finest B-school, I build tech powered solutions to ensure maximum profitability for any business venture.
Skills :
● Data Science & Machine Learning : Supervised ML, Linear Regression, Decision Tree, Prompt Engineering, Hypothesis Testing, Statistics, Exploratory Data Analysis, Data Wrangling, Cleaning & Tidying, Feature Engineering, Data Visualization, SQL, Scikit-learn, NumPy, Pandas, Matplotlib, Plotly, Dash, Microsoft Excel, Metabase, AWS Athena & Sagemaker, Google BigQuery, Unsupervised ML, Neural Nets, NLP, Large Language Models (LLM), GPT, Text Embedding, Vector Databases, PyTorch, fastai, TensorFlow, OpenAI API, Generative AI (GenAI)
● Programming : Python, MySQL, PostgreSQL, HTML, CSS, Flask, Heroku, APIs, Git, AWS S3 & EC2, JavaScript, jQuery, Java, Werkzeug WSGI, Leaflet.js, Folium, GeoPandas, Google Maps API & Sheets API, Bootstrap
Professional Work :
Engineered a suite of cutting-edge ensemble ML models, revolutionizing underwriting processes for personal loan products a major FinTech Lender in India.
Developed and deployed 6 ML models for credit-risk scorecard, using 5 data sources (Bureau, Bank Statement, SMS, Smartphone App Info and Demography), improving risk spread by 1200%.
Devised NLP based models (Passive Aggressive Classifier and XGBoost Classifier with TFiDF Vectorizer) for classification of SMS, improving recall by 35% and accuracy by 60%, and saved 36 man-hours daily, processing over 200 Mn SMS/day.
Created LLM based classifier models using Prompt Engineering on ChatGPT (OpenAI), Mistral AI (Open Source) and other open source models. Deployed using AWS Sagemaker.
Crafted an LLM based Law ChatBot that answers to Lawyer's questions pertaining to any specified PDF file of a Legal Document.
Built an in-house MLOps infrastructure with AWS Sagemaker, enhancing model integration.
Engineered 330+ features from SMS data to emulate metrics including cash flow, lifestyle and financial literacy indicators. Included steps like Exploratory Data Analysis, Data Wrangling, Cleaning & Tidying, Feature Engineering.
Designed comprehensive daily and monthly model monitoring dashboards using Flask-Python, Metabase and SQL tables, improved response time to anomalies by 75%, enhancing overall model reliability and performance.
Designed a Customer Loyalty Program and delivered via a web and Android app, over 2500 DAU, rated 4.6 on play store for a manufacturer and distributor of PVC Plumbing Systems in Central India.
Increased territorial penetration by over 25% by building automated interactive GIS sales dashboard for a wholesale business in India.
Achieved 100% monthly revenue target completion by creating Business Intelligence Dashboard – Google Apps Script.
Reduced total credit by 20% in five months by building a Python-bot that sends regular WhatsApp messages to debtors.
Saved annual recurring cost of ₹25 lakh+ by automating data collection for a hedge fund using Yahoo Finance API & Python.
Cut time from over 4 hours to 2 clicks by developing a GIS-enabled web application for new-license territory discovery for a developer and licensor of Escape Room Game Designs in USA.
Personal Projects :
Developed a Fake News Classifier using TfIdf Vectorizer and Passive Aggressive Classifier - ML, pandas, scikit-learn.
Created a Spam Email Classifier using Naïve Bayes and XGBoost Classifiers - ML, NumPy, seaborn.
Designed a Cat Image Classifier using multi-layer neural network - DL, NumPy, Matplotlib, h5py, PIL, SciPy.
Developed a web app to manage portfolio of stock with real-time prices using IEX API -SQLite, Flask, HTML, CSS, JS.
Built a terminal-run card game called Declare - Python.
Education :
Computer Science and Data Science Masters (Self-Learned)
Major Electives:
Computer Science: MIT 6.0001, MIT 6.0002, Automate the Boring Stuff with Python, Harvard CS50, Harvard CS50 - Web Development, Intro. to Algorithms, Design and Analysis of Algorithms, Software Development Lifecycle, GIT and GitHub, Flask for Python, The Python Mega Course, Data Structures, Analysis of Algorithms, The Odin Project etc.
Data Science: Stanford CS229, Stanford Machine Learning (CS229A), Data Science path on Codecademy.com, Data Science and Machine Learning Bootcamp, Neural Networks and Deep Learning, Deep Learning with PyTorch: Zero to GANs, Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization
Data Scientist and Machine Learning Engineer specializing in building production-grade ML systems across diverse domains including manufacturing optimization, anomaly detection, computer vision, NLP, and geospatial analytics. At Novelis Inc., I develop scalable, real-time ML pipelines integrating optimization techniques, computer vision, and AutoML methods deployed on cloud-native infrastructure.
Previously at the Ministry of Rural Development, I designed large-scale NLP and geospatial clustering models deployed at a national scale to optimize infrastructure planning and operational efficiency. My past experience also includes developing deep learning-based energy optimization solutions at ActiveBAS and delivering production ML services at scale, such as an air quality forecasting API at Blue Sky Analytics.
My academic research at Michigan State University focuses on advanced techniques, including Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), Graph Neural Networks, and Semantic Segmentation. I have hands-on experience deploying models in cloud environments (AWS, Azure, GCP), alongside strong proficiency with Python, PyTorch, TensorFlow, SQL/NoSQL, and containerized applications.
Open to full-time roles as a Machine Learning Engineer, Applied Scientist, or Data Scientist in organizations tackling complex AI-driven challenges at scale.
