Yash Khemchandani Email & Phone Number

Bengaluru, Karnataka, India

Gyeonggi, South Korea

- Lead developer for building containerized microservices (xApps) for RAN Intelligence Controller to improve the 5G network quality- Designed and implemented algorithms to dynamically allocate network resources using performance metrics like Packet Delay and Throughput- Built a scalable and fault‐tolerant architecture for handling massive user traffic without compromising the network service requirements- Collaborated with engineering teams from 4 countries to develop the commercial packages for Telecom operators including KDDI and DISH

Remote

- Developed low-latency infrastructure for buying personal loans on P2P lending platforms using ML-based loan default predictions- Implemented batch processing to efficiently handle large datasets and significantly reduce model training time- Redesigned the existing database schema and handled data migration, to allow scaling to multiple platforms and asset classes- Built a simulation environment for back-testing models on historical data

Mumbai, Maharashtra, India

- Worked under Prof. Sunita Sarawagi in collaboration with Google Research India to extend BERT-based language models to Indian languages with limited corpora and different scripts- Exploited the relatedness between low‐resource Indian languages and Hindi to augment training data and improve performance- Gained experience in training large models on textual data using cloud TPUs

Seoul, South Korea

- Implemented state-of-the-art Object Detection models to discover 5G radio antennas on rooftops and detect any anomaly in their installation- Used Transfer Learning to achieve good precision despite small dataset

Canberra, Australian Capital Territory, Australia

- Designed and implemented U-Net based Image Segmentation models to locate neck muscles in MRI Images for diagnosing conditions like whiplash

Liverpool, England, United Kingdom

- Implemented a Reinforcement Learning pipeline to generate novel and chemically valid molecular structures with desired properties- Trained Graph Convolutional models for predicting the properties of molecules, and used them as a reward for the RL pipeline

Mumbai, Maharashtra, India

- Conceptualized several competitions related to Literary Arts, Speaking and Quizzing- Responsible for finding sponsors and reputed judges for the competitions in order to encourage greater participation

B.Tech (With Honors), Computer Science, 9.63/10

Minor in Applied Statistics and Informatics Minor in Artificial Intelligence & Data Science Teaching Assitant: - DS303 Introduction.

High School

Intermediate : 97.6% Matriculation: 10.00/10.00 Member of Quizzing Club. Won city round of Bournvita Quiz Contest and Aqua Regia.