Chaitanya K. Joshi

Hi, I'm Chaitanya Joshi! I'm a Research Engineer at A*STAR, Singapore, working on the emerging field of Graph Deep Learning. I graduated as the Valedictorian from Nanyang Technological University, Singapore in 2019 with a B.Eng. in Computer Science, and was a Research Assistant under Prof. Xavier Bresson.

Research Interests

My research focuses on Graph Neural Networks (GNNs) and is motivated by their application to accelerating scientific discovery. Towards developing the next generation of GNN architectures, I have open-sourced better benchmarks for modern GNNs and am exploring their theoretical and empirical limitations.

A key application area motivating my work is the intersection of Deep Learning and Combinatorial Optimization, in particular the famous Travelling Salesman Probelem (TSP). My papers on deep GNNs for learning-driven TSP solvers and the challenge of out-of-distribution generalization for GNNs have been presented as an invited talk at INFORMS 2019 and a workshop poster at NeurIPS 2019. I was also awarded NTU's Best Final Year Thesis for my work.

Previously, I have worked on SAP's flagship enterprise ML product, Cash Application, co-designing a deep learning system for matching financial data to automate corporate accounting processes. Today, CashApp handles over € 200 Million Euro in annual sales pipeline for global companies and lead to 3 US patents.

Before that, my journey in Machine Learning started with Natural Language Processing research at École Polytechnique Fédérale de Lausanne and deploying Dialog Systems at conversational AI startup Evie.ai.

Technical Writing

In addition to research, I am deeply passionate about communicating technical ideas in accessible digital formats for both research audiences and the general public. Most recently, I published a popular article bringing together my interests in GNNs and NLP with The Gradient.

Researcher @chaitjo describes how the popular Transformer architecture is actually a Graph Neural Network in disguise! This suggests many potential areas of research where NLP can learn from graphs! #nlp #graphshttps://t.co/FEnKxJIZc7

— The Gradient (@gradientpub) September 12, 2020

Here are some recent highlights from my blog: