• Fast-moving trends rapid changes in the business environments, and concomitant changes in consumer behavior, makes it critical to track the demand closely especially in the short-term horizon.Thoucentric Demand Sensing platform (DSP) delivers accurate short-term demand forecast leveraging inputs from disparate systems such as endogenous & exogenous demand drivers.
• Improved short-term demand forecast accuracy by ~30-40%.
• Increased inventory accuracy by up to 20% leading to optimal inventory levels.
• Architected and built a scalable and reproducible Data Pipeline that collects, validates, analyzes, and preprocess data from 30+ sources and creates 50+ features and outputs a model ingestible DataFrame for sensing demand.
• Developed a production-ready multivariate modeling pipeline from multiple Jupyter Notebook experiments done by Data Scientists.
• Created a COVID-19 engine with 50+ features extracted from Depot and Channel level COVID-19 related information, to forecast a multiplier that can adjust the forecast from the Demand Sensing model among this pandemic situation.
• Refactored and optimized multiple Machine Learning codebases.
• Tech Stack: Python3+, Azure, Pandas, scikit-learn, fast.ai, statsmodels, Git

• Website: https://www.planetspark.in/
• Google playstore link: https://play.google.com/store/apps/details?id=in.planetspark.cosmos
• Created APIs and background jobs for score, accuracy, leaderboard generation for games.
• Developed trend-detection logic for finding trending games.
• Built keyword based image retrieval job for finding game images based on game title and description.
• Made a recommendation system for realtime game recommendations.
• Documented the design workflow for the whole codebase with ER diagrams.
• Used technologies: Ruby on Rails, Postman,SideKiq, Redis, Git/Gitlab, AWS

This was a summer School on "Theoretical Foundations of Computer Science" organized by IIIT Bangalore and sponsored by Sonata Software. This was a resume based selection across the country. They offered tutorials on Approximation Algorithms, Parameterised Algorithms and Complexity, Cryptography, Program Analysis and Formal Methods, and Theoretical Foundations of Distributed Computing. Each of the topics was covered briefly. The speakers were from IIIT Bangalore, IBM Watson, Microsoft Research and Sonata Software.

Worked under the team of T.C.O.Developed a software for taking photos of answer scripts with two webcameras simultaneously by just turning the pages and then convert them into a single pdf for an individual student, so that a student can check his/her answersheet on ERP.

Deep Learning based statistical downscaling and future projection of temperature in Bhuntar, Himachal Pradesh. The main objective of this project was to compare the results of Machine Learning, SDSM, ANN and nonlinear regression method for statistical downscaling and future projection of temperature.

Built an object detection software which is able to detect class of all the objects present in an image served as an input or in real time. Convolutional Neural Network, AWS Rekognition, Microsoft Vision, Tensorflow Object Detection, OpenCv, YOLO, Python, Tkinter have been used to build the software and Git/Github, Jenkins have been used for build automation and Docker for deployment.

Developed a blogging app in Ruby on Rails where user can post article category wise and view other user's article.

Developed a stock tracking social media application using Ruby on Rails where users can track stock using ticker and add in their tracking list.Also users can follow other users and add them as a friend to check what they are tracking.

Developed a GUI based Java code editor using Java Swing which has functionalities like Autocomplete,Syntax highlighting etc..

This project is to study the economic growth and the market for the semiconductor industries in India.My role was to analize the impact of technology and innovation in the semiconductor industries in India.

This was a research project on the following topic- Can nanotubes be used in place of copper cables for reducing transmission line losses as they have lower resistance in comparison with copper cables? Can carbon nanotube wires be used in high-frequency transformers and inductors?