Karan Tanwar

AI Researcher

New Delhi, Delhi, India6 yrs 7 mos experience

Most Likely To SwitchAI ML Practitioner

Key Highlights

Expert in deploying scalable AI solutions.
Proven track record in machine learning projects at Amazon.
Strong academic foundation from IIT Delhi.

Stackforce AI infers this person is a Machine Learning Engineer with a focus on AI-driven solutions in E-commerce.

Contact

Skills

Core Skills

Machine LearningFraud DetectionSoftware DevelopmentCloud Computing

Other Skills

AWS S3AlgorithmsAmazon DynamoDBAmazon ECSAmazon Elastic Container Registry (ECR)Amazon Web Services (AWS)Artificial Intelligence (AI)Artificial Neural NetworksCommunicationComputer NetworkingComputer VisionDatabasesDigital Image ProcessingFeature EngineeringFull-Stack Development

About

Meesho’s Fraud AI team leverages advanced machine learning techniques to enhance fraud detection capabilities, where contributions focus on deploying scalable solutions using PyTorch and PySpark. Previous experience as a Software Development Engineer at Amazon honed expertise in international machine learning projects, emphasizing cross-functional collaboration and technical precision. Graduated from IIT Delhi with a dual degree in Computer Science and Engineering, combining academic rigor with hands-on experience such as serving as a Graduate Teaching Assistant for a machine learning course. Motivated by solving complex problems, the aim is to develop impactful AI-driven solutions that address real-world challenges.

Experience

Meesho

Data Scientist

Apr 2024 – Present · 1 yr 11 mos

Fraud AI Team

PyTorchPySparkMachine LearningFraud Detection

Amazon

Software Development Engineer

Jul 2022 – Apr 2024 · 1 yr 9 mos · Bengaluru, Karnataka, India

Worked as a part of International Machine Learning team

JavaLarge Language Models (LLM)PyTorchAmazon ECSPySparkAmazon Elastic Container Registry (ECR)+2

Indian institute of technology, delhi

Graduate Teaching Assistant

Aug 2021 – Dec 2021 · 4 mos

TA for the course - Fundamentals of Machine Learning (COL341) Fall 2021, under prof. Rahul Garg.

Freecharge

Software Engineer Intern

May 2021 – Jul 2021 · 2 mos · Mumbai, Maharashtra, India

Digital Business and Transformation (DBAT) Team
Developed APIs, which provides access to AWS S3 files. Used aws-sdk to create a filesystem which can delete, list, and give presigned url to manage data (directly from S3 bucket). Reduced the server load by 20%.
Used MongoDB to cache all the file paths and meta data such as access permissions for each file in S3 buckets.
Received pre placement offer.

Amazon

Applied Scientist (ML)

Feb 2021 – May 2021 · 3 mos · Bengaluru, Karnataka, India

Team: IN Machine Learning | Project Title: Search by Data
Developed unique algorithm which, for any dataset with 100+ attributes (such as item name, brand name, item weight), recommends top-k features which will result in increased accuracy/PR-AUC/ROC-AUC in classification problems.
Before this project, either only item name was used, or attributes were manually selected (for ex using product material attribute for classifying smartphones). Using all features has a high computational cost, and most of the times, does not give the best models.
Used Vowpal Wabbit for faster training and explainability for ranking feature importance (text and categorical).
Showed a substantial increase in ROCAUC and PRAUC from baseline results for AB (Amazon Business) datasets.
Received pre placement offer.

Vowpal WabbitMachine Learning

Board for student welfare (bsw), iit delhi

Student Mentor

Aug 2019 – Feb 2021 · 1 yr 6 mos · New Delhi Area, India

Mentoring seven Freshers to get settled at IIT in terms of academic and social life.

Rendezvous iitd

2 roles

Technical Coordinator

Jun 2019 – Oct 2019 · 4 mos · New Delhi Area, India

Worked as a technical head, supervising 12 Activity heads, for the backend programming for the website of IIT Delhi’s cultural Fest, Rendezvous.
Created Admin and Sponsors portal, deployed on Amazon AWS using Elastic Beanstalk.
Created general pass system that allow users (students and faculties) to register or participate in an event.