Arut Selvan Dhanasekaran

Arut Selvan Dhanasekaran

Software Engineer at WhatsApp | Distributed Systems & AI

Experience

Software Engineer | WhatsApp (Meta) Feb 2024 – Present

Owner of contacts infrastructure for 3B+ users. Optimized privacy checks from 4M/day failures to <100/day. Building discovery APIs for Meta AI bots.

Software Engineer | ServiceNow May 2022 – Feb 2024

Led development of high-performance autocompletion features. Leveraged Rust-based SWC transpilers to optimize JS execution in JRE.

Software Engineer II | JPMorgan Chase & Co 2019 – 2021

Engineered unified monitoring for firm-wide encryption infrastructure using Python, Java, and ReactJS.

Research Work

Benchmarking Generalization via In-Context Instructions

Co-authored seminal work on instruction tuning used as foundational training data for FLAN-T5, OPT-IML, and Stanford Alpaca.

1,245+ Citations Scholar Link →

Detection of Pitting Corrosion using RCNN

Developed an automated framework using Regional CNNs to detect and analyze corrosion in metal surfaces. This Deep Learning solution enables predictive maintenance for industrial structural integrity.

Deep Learning / Computer Vision GitHub →

Projects

Bloom

NLP engine based on Bloom’s taxonomy. Awarded by the Prime Minister of India at Singapore-India Hackathon.

View Repo

natural-instructions-expansion

Expansion of the Natural Instructions dataset for cross-task generalization in LLMs.

View Repo

selective_prediction_mtl

Researching sampling strategies for Selective Prediction performance in Multi-Task Learning.

View Repo

tweets_sentiment_analyzer

Real-time sentiment analysis engine for live Twitter streams using NLP and Python.

View Repo

react-fullscreen-slideshow

Published NPM package for high-performance React slideshow components in TypeScript.

View Repo

Awards

2022

1st Place

ServiceNow Intern Hackathon (1/40 teams)

2018

2nd Place

Singapore India Hackathon. Awarded by PM of India (2/40 teams)

2017

1st Place

Software AG API Hackathon (1/30 teams)