Hi!
I'm Rajdeep Biswas, a Data Scientist with experience in engineering intelligent systems.
Get in touch r4jdeepbiswas@gmail.com
I'm currently a Data Scientist & Machine Learning Engineer at SAP Labs, Bangalore building Machine / Deep Learning systems from scratch with some awesome people.
I am a musician turned programmer.
(Also, a metalhead!). 
I started getting into programming languages with the belief of the power of being able to talk to computers will let me ease up a lot of things, and it just snowballed from there.
I discovered machine learning in my sophomore year of college and I haven't gotten over the idea of making AGI possible, since.
As a side effect of trying to understand machine intelligence, I find myself deeply intrigued by the underlying workings of human intelligence.
I can't wait to find some time for myself (hopefully, in the near future) to study some Neuroscience (and brush up my Calculus, too).
I am also a philosophy & psychology nut, and a massive Rick and Morty fan.
- Python
- Java
- SQL
- Flask
- TensorFlow
- PyTorch
- Sklearn
- Pandas
- Numpy
- Matplotlib
- Git & Github
- Jupyter Notebook
- Bash
- Postman
Most of the code that I have written stays in Jupyter Notebooks or my company's proprietary deployed applications. Below are linked some projects that I can show you which best showcase the structure and documentation of the code that I write.
Complete automated analysis solution to parse your invoices that makes text preprocessing recommendations to enrich your UNSPSC classifications better.
A ton of helpful Time Series data preprocessing utility I wrote for an Anomaly Detection (LSTM - Autoencoders) Service on incoming API Traffic usecase.
Takes in an Excel Sheet containing commodities and their expenditure histories, automatically connects to a remote HANA Db runtime via a python wrapper, performs iterative analysis of all available forecasting algorithms, generates & saves comprehensive plots of the predictions, and at the end generates a report of the best performing algorithms.
A Linear Regression model with a Polynomial features of degree 4 learn the data of last 30 days of growth of COVID-19 number of cases (cumulative) and try to predict what the following 15 days would look like. View live graph here.