Prajwal Singhania

Associate Consultant

College Park, Maryland, United States7 yrs 11 mos experience

Highly Stable

Key Highlights

PhD student focused on AI training optimization.
Top of class graduate from IIT Kharagpur.
Experience in developing low-latency trading systems.

Stackforce AI infers this person is a skilled AI researcher with a strong background in Fintech and High Performance Computing.

Contact

Skills

Core Skills

High Performance ComputingDeep LearningSoftware Development

Other Skills

BashCC++CSSCUDACustomer ServiceELK stackHTMLInternational RelationsJavaScriptKibanaLaTeXLeadershipManagementMicrosoft Word

About

I am a CS PhD student at the University of Maryland - College Park, working under Dr. Abhinav Bhatele. am interested in developing and optimizing systems for AI training and inference. Currently, my work focuses on scaling parallel AI training frameworks to a large number of GPUs and developing faster, memory-efficient LLM inference techniques I graduated top of my class from IIT Kharagpur in 2020 with a Dual Degree (B.Tech. + M.Tech.) in Computer Science and Engineering. Before joining my PhD, I was a team lead (Senior Assoicate - Trading Systems) at AlphaGrep Securities, developing low-latency trading systems for Crypto and Indian Exchanges.

Experience

7 yrs 11 mos

Total Experience

3 yrs 11 mos

Average Tenure

Current Experience

Microsoft

Research Intern

Jun 2024 – Aug 2024 · 2 mos · Bengaluru, Karnataka, India · On-site

Deep LearningHigh Performance ComputingResearch

University of maryland

Graduate Assistant

Aug 2023 – Present · 2 yrs 10 mos · College Park, Maryland, United States · On-site

CUDAHigh Performance Computing

Alphagrep

4 roles

Senior Associate / Team Lead

Promoted

Jul 2023 – Aug 2023 · 1 mo · Bengaluru, Karnataka, India

Associate

Jul 2022 – Jul 2023 · 1 yr · Bengaluru, Karnataka, India

Senior Analyst, Trading Systems

Jul 2021 – Jul 2022 · 1 yr · Bengaluru, Karnataka, India

Analyst, Trading Systems

Jul 2020 – Jul 2021 · 1 yr · Bengaluru, Karnataka, India

Tower research capital

Risk Technology Intern

May 2019 – Jul 2019 · 2 mos · Gurgaon, Haryana, India

During my internship at Tower Research Capital Pvt. Ltd. (Gurgaon, India), I completed the following tasks:
Set up a pipeline for simulation of an existing monitoring tool on historical dates which will help configure better parameters for the tool in the future
Deployed an end-to-end product for the same; developed the backend using C++, Bash, Python (Flask) and the UI using HTML, CSS, JavaScript
Incorporated features like higher replay rate for the simulation tool and better log readability (critical for building up analytics) into existing codebases
Used the ELK stack to index the logs of the monitoring tool and create visualizations for log analysis on Kibana, which is critical for risk assessment

Usc viterbi school of engineering

Summer Research Intern

May 2018 – Jul 2018 · 2 mos · Los Angeles, United States of America

As one of the 18 people across India to be selected for IUSSTF-Viterbi Summer Research Program, I spent my summer of 2018 working under the guidance of Prof. Meisam Razaviyayn and did the following:
A literature study of previous works related to Quantization and Binarization of Neural Networks
Came up with the idea of adaptively quantizing different layers of a neural network with different number of bits
Modified existing Tensorflow and Pytorch models to test out various hypothesis to improve accuracy
Studied the effect of different regularizers to aid in the quantization/binarization of the layers of a neural network
Developed an algorithm using regularization to optimally choose the quantization for each layer during training

Industrial technology research institute (itri)(工業技術研究院, 工研院)

Summer Intern

May 2017 – Jul 2017 · 2 mos · Hsinchu County/City, Taiwan

As a summer intern in the Information and Communications Lab of Taiwan's most prestigious R&D Institute, I did the following:
Worked on a "Musical Style" Recognizer to classify music into the genres: rock, jazz, classical, gospel, rap
Trained the model on our own dataset of 5000 musical pieces(of 1-minute duration each) extracted from Youtube
Used the weights of the trained CNN classifier to implement Neural Style Transfer on the pre-processed inputs (spectrograms from mp3 files) and get a musical piece with the desired melody in the desired style