Full Stack Data Scientist & Business Systems Developer · Finance background · ML, Automation, Django
I build data and software systems for businesses, and I came to this work from the business side.
Before I wrote production code, I worked in accounts: bookkeeping, ledgers, reconciliations and the profit-and-loss statement at month end. Before that, I taught Tally ERP and computer accounting to students preparing for their first jobs. That background shapes everything I build. When an owner says "sales are up but cash is short", I know which reports to open, where the numbers usually go wrong, and what a model or an automation can realistically fix.
Today I work as a Data Scientist on live business problems: a product recommendation system and website for a food & beverages company, and data cleaning, reporting and automation for a textile business. I also run BeeProton AI, where I design and build websites, client portals, GST-compliant billing, automations and machine-learning solutions for small and growing businesses.
I learned most of this by doing it, on real books of accounts, real sales data and real deadlines. Certificates from IBM and Harvard gave me structure. The work taught me what matters: a solution is finished only when it saves time, reduces errors or helps someone make a better decision where money is at stake, and when we can show that it did.
Click any role to read what the work actually involved.
Running a technology services business end to end: scoping, fixed-price proposals, delivery, invoicing, collections and after-launch support.
Data and digital work for an established food & beverages brand.
Turning a growing business's day-to-day data into something management could rely on.
Day-to-day accounting, with data analysis applied to the books.
Job-readiness training in computers and accounting.
Filter by area, or search for something specific.
Nothing matches that.
Business systems, websites, machine learning and data work. Each case study starts with the business problem and ends with how the result is measured. Source code is on GitHub where it is mine to share.
Brijwasi Sweets (India) Pvt. Ltd.
A new company website and an AI product-recommendation system for a food & beverages brand, aimed at larger orders and more repeat customers.
BeeProton AI (in-house)
The system BeeProton AI runs on: client portal, GST-compliant invoicing, payments, milestones, support tickets and encrypted client communication, in one place.
Anvi Special Education and Rehabilitation Foundation
A ten-page website for an NGO in Navi Mumbai, built around one goal: helping parents book an assessment. Includes UPI and bank donations with no gateway fees.
BeeProton AI (in-house)
Privacy-first PDF, image, Excel and data tools that run entirely in the browser, plus 16 free course tracks: a marketing asset that earns traffic instead of buying it.
Time-series forecasting of sales from historical data, built for purchase, stock and cash-flow planning.
A classification model that ranks customers by how likely they are to leave, so retention effort goes where it pays back.
An NLP recommender that suggests similar items from their descriptions: the technique behind "you may also like", demonstrated on a movie catalogue.
An intent-classification chatbot that answers common questions in English and Hindi, so staff spend less time on repeat queries.
A deep-learning image classifier for 9 skin conditions with a Streamlit app: a computer vision project that also taught when not to trust a model.
Fine-tuning pre-trained CNNs (ResNet / VGG / AlexNet) on a small image set: how to get a useful model when labelled data is scarce and expensive.
No project matches that.
Tell me what you are trying to decide, measure or automate, and what it is costing you today. If I am not the right person for it, I will say so.
