About Me

My dream is to use ML Techniques to make programming systems safer and easier to build and maintain, and to take inspiration from formal methods to make machine learning more reliable. I have only begun on this journey, and I am still understanding the fundamentals of these areas. If you have something to teach me, I’d love to learn.

Prior to my stints at Microsoft Research working on AI agents and models for large enterprise-grade software engineering, and at Georgia Tech on Probabilistic Model Checking, I have worked as a Platform Engineer at Chorus One on the Infrastructure Team, where I built tooling for secure key management and encryption of their servers. I have completed internships at Adobe Research as a Research Intern and DataChannel Technologies as a Data Analyst Intern. I have worked with Prof. Subodh Sharma in a project on the formal verification of concurrent programs, and later on a COVID Vaccine Alerts telegram bot which reached people from 400+ districts.

I took lots of interesting courses at IITD, including Computational Neuroscience, Automated Verification of Concurrent Programs, Semantics of Programming Languages (where I wrote a Lean proof!) and Program Synthesis for Smart Contracts. The most interesting one remains Proofs and Types, where I enjoyed reading old (and poorly typeset!) papers by Church, Kleene, Rosser and the like. I also write these notes I’m very proud of: Lecture Notes on λ-definability ≡ Recursiveness, the Scott-Curry Theorem (Undecidability of β-equality in the untyped λ-Calculus). You can find more miscellanous slides and writings at https://ramneet-singh.netlify.app/publications/.

In the Summer of 2022, I worked as a Research Intern at the Big Data Experience Labs, Adobe Research. We worked on making segments easier to understand, analyze and predict for marketers. On the technical side, we modelled market segments using temporal graphs, and then applied neural embedding techniques for temporal graphs to the tasks of edge-weight and node-feature prediction. This enabled us to predict how customers will flow between segments over time, allowing marketers to understand their market better and make better strategies.

In the Summer of 2021, I interned as a Data Analyst at DataChannel Technologies. We worked on using historical marketing data across various media/campaigns/adsets to predict their performance curves, and subsequently used those to optimize the marketing budget distribution among them, also incorporating an option to impose typical marketing constraints.

During my second year, I was involved in a research project on the formal verification of concurrent programs under Prof. Subodh Sharma. We implemented a novel algorithm for value-equivalence based model checking, adding a custom tool to a large open-source software (Nidhugg) used for finding bugs in concurrent programs.

At the end of my first year, I led a team of 3 to secure Third Place in the HCL Hack IITK, a cybersecurity hackathon organised by C3i Hub IIT Kanpur. We built exciting Machine Learning based solutions to real-world cybersecurity problems. Check out our code or our finals presentation.

With an aim of contributing to the ML Community at my college, I co-founded the AI/ML Club at IIT Delhi. I also created the Co-WIN Vaccine Alerts Telegram Bot, which helped people from 400+ districts across India find vaccination slots. I can watch/play/talk about sports all day, and love playing badminton. Hit me up for stories from my 1st Year of how we (Udaigiri hostel) won in badminton (in an epic final by the way) and came third in squash.