About
A keen learner, passionate about cutting-edge technology and solving real-world problems through Big Data and Deep Learning who love building digital products.
Data Scientist & Machine Learning Engineer.
Your Data has Powers, I harness it.
- Name: Rishabh Vij
- Website: rishabh-vij.github.io
- City: City : Mumbai, MH
- Currently: Data & AI @ Paytm Insider
- Email (Work): rishabh@insider.in
- Email(Personal): rishabhvij31@gmail.com
Skills
I'm a seasoned Data Scientist and ML Engineer who loves building AI powered products.
Resume
Summary
Rishabh Vij
I'm a data-driven Data Scientist with 2+ years of experience designing and developing AI powered products from initial concept to final, polished deliverable.
- Mumbai, MH, India
- rishabhvij31@gmail.com
Education
Bachelor of Computer Applications
2015 - 2018
Symbiosis International University, Pune
- Computer Fundamentals, Operating Systems
- Data Structures, Relational and Non-Relational Databases
- Linear Algebra, Statastical Mathematics
- Data Mining, Data Warehousing, Big Data
- Human Computer Interaction, Product Design, Cloud Computing, IOT
Professional Experience
Machine Learning Engineer
2018 - Present
Paytm Insider, Mumbai, MH
- Designed and Developed ALS powered Recommendation Engine, Developed Dynamic Pricing service for revenue optimization.
- Developed AI powered services like automated ID verification and Dynamic Image Background Removal for PayTM insider Accreditation.
- Designing and Developing data pipelines and ML powered services for PayTM Movies and PayTM Feed, which are both high scale systems.
Computer Vision Developer
2018 - 2018
Sensus Labs, Mumbai, MH
- Designed and developed deep learning powered service to serve in store analytics which are generated post processing in-store CCTV stream.
- Developed a service which could automatically prepare and update a dataset of customers via CCTV Feed.
Projects
In my free time I often end up exploring open source dataset for ML modelling or Kaggling. Some of my cool side projects are:
Art Generation - Neural Style transfer
Given a style and a content image, this model can generate novel art
World Dog Breed Finder
Given an Image of a dog, this model can tell which one out of the 120 breeds known to it does the image belong to
Face Verification
Inputs an Image and if the image is verified as one the 5 people in the dataset, displays the code word
MovieLens Recommendation System
A content based and collabarative filtering based approach to the MovieLens dataset