Everyone is waiting for the right moment.
There was no moment. I made one.

A Little About Me
- I'm a Data & AI Engineer who likes turning messy data into software people actually use.
- I work across software engineering, data analytics, and machine learning, building interfaces and APIs that stay fast.
- On the AI side, I integrate LangChain, Gemini, and Ollama into production workflows — RAG pipelines, agents, and self-correcting validation loops.
- Always up for a chat, so feel free to reach out.
A Few Things I Love...
Places I've worked
Product Intern
/ Atypical Advantage- Developed and integrated a multilingual STT/TTS application supporting all 22 official Indian languages with automatic mid-conversation language detection.
- Curated, structured, and preprocessed datasets for ML-based conversational learning systems, supporting reliable downstream model development.
- Contributed to the application structure for an accessible, inclusive user experience.
Research Intern
/ Malaviya National Institute of Technology (MNIT) Jaipur- Developed an end-to-end ML pipeline to detect stealthy cyberattacks on power grids by simulating adversarial manipulations of load data.
- Used clustering to baseline normal load patterns and define parameters for stealthy anomalies.
- Engineered a high-dimensional feature set (power flow measurements, statistical dispersion indicators) and trained a classifier to separate normal operation from targeted false data injections.
Research Intern
/ Indian Institute of Information Technology (IIIT) Allahabad- Cleaned and processed a combined dataset of 60,400+ records with Python and Pandas to train and validate transformer models (MentalBERT, MelBERT) for mental-health sentiment analysis.
- Analysed misclassification patterns to build an optimised two-stage hierarchical classifier.
Things I Know
View CertificationsThings I've Built
View all projects
Demand Forecast & Reorder Simulation
completedSKU-level 7-day demand forecasting for 100 retail SKUs with LightGBM, tested against simple baselines with leakage checks, and a simulated reorder policy that shows what better forecasts do to stock and fill rate.

Instacart Customer Analytics
completedSQL and Python analysis of 3.4M Instacart orders covering the funnel, retention, reorder behaviour and basket affinity, plus a simulated A/B test whose design and decision rule were fixed before running anything.
Things I've Published
Multivariate Deep LSTM for Short-Term Load Forecasting: A Robust Framework for Delhi Under Extreme Weather
Short-term electricity load forecasting for Delhi with deep LSTM networks, focused on how the models behave during extreme weather.
Distilling Black-Box Ensembles into Interpretable Decision Trees for Explainable AI
Distils the behaviour of black-box ensemble models into decision trees that people can read, so predictions stay explainable.
Let's Talk
Words that stayed
Everyone is waiting for the right moment.
There was no moment. I made one.

