About me

I am in my fifth (and final) year of a PhD in Operations Research, advised by Yuri Faenza at Columbia. I am an engineer, mathematical modeler & researcher, and software architect. I am graduating in mid-2025 and am currently looking for work in New York City in finance or software engineering.

PhD Research

My research is on stable matching and combinatorial optimization, largely motivated by public school choice in very large school districts (such as the one in New York with 1.1 million students). The tentative title for my thesis is “Tradeoffs between Information, Tractability, and Fairness in Large Matching Markets”. I have found success in applying concepts from probability and analysis into this field to prove a number of interesting results.

Stable matching is one of the cornerstones of modern market design, and is therefore heavily studied by both the operations research and economics communities. Such models arise whenever a central planner is tasked with pairing together agents of two distinct types, with each agent having idiosyncratic preferences over the agents of the other type, and money cannot be used to clear the market. Natural applications are found in the presence of indivisible goods when money is excluded due to fairness considerations; for example in refugee settlement, medical residency matching, or the allocation of public goods like school seats or public housing.

In large markets like New York school choice, there are inherent difficulties in providing full information. For example, it’s completely infeasible for students to list out their full preferences among all programs in the city, which number in over a thousand. Likewise, when schools are allowed to fully define their preferences (via “choice functions”) and indicate preferences over different assembled classes, it becomes completely infeasible to describe or even solicit from schools these full preference systems. Most real world implementation therefore restrict participants to describe their preferences using only lists (often truncated), which significantly impairs the abilty of the agents to describe their true preferences.

My research is focussed on studying this information gap, and finding practical methods to move from the current limited-information setting closer into the full-information setting, while still maintaining feasibility of the algorithms and system mechanisms.

Other activities

In addition to and concurrently to my PhD, I am active in a number of other endeavours. Below is a sampling of some projects I have been involved with recently.

Couchers.org: a non-profit, open-source couch surfing community

I co-founded Couchers.org, a free and open source couch surfing platform and community. We have well over 50 000 users, distributed across the entire globe. Couchers is focussed on being a safe, healthy and welcoming community, and in particular welcoming newcomers and those who might not traditionally feel comfortable with the concept of couch surfing.

I currently lead the project through active involvement in strategy and engineering as well as being the President of the Board of Couchers, Inc, the United States-based 501(c)3 tax-exempt non-profit we are incorporated as.

If you don’t know what couch surfing is, here’s my 2 minutes pitch. We all love travelling to visit friends in new cities: it lets us see the place from a local’s opinionated point of view — what they like and what real life is like in their city. It gives us a view into the authentic experience beyond the touristy stuff. Couch surfing is like this, but you also make that friend as you do it. You search for a host that you think you might get along well with, anyone who has put up their “couch” (though most often people have a spare bedroom) up on the site, then send them a request and arrange to stay at their place.

I find that the couch surfers I meet on Couchers are surprisingly likeminded in some ways but often drastically different. Sharing a roof and meals with them, and the intense connection that is formed in these few-day interactions gives you a deep appreciation and insight into someone else’s life in a way that few other interactions can. I believe that this is extremely powerful for our ability to relate and empathize with others, and I particularly believe being able to facilitate these connections for young people is meaningful for them and helps them grow as people (like I did through my many travels).

I have heard many amazing stories of empathy and raw human connection and the positive effects from them from both men and women on the platform, and that’s what keeps me motivated to continue working on it.

Ask2.ai: bringing AI and ML to everyone in finance

Ask2.ai provides a platform and tailored solutions to various financial clients seeking to apply AI, machine learning, and LLMs in the financial field. Ask2 is both a research organization — with a surprisingly large research output for its size (all published openly) — as well as a cutting edge engineering team implementing these in-house developed models plus the industry leading algorithms under one unified interface. The platform allows power users to interact with it via a powerful API, but also enables easy access to data and analyses via an easy-to-use web-based interface. These underlying “engines” can be applied to the whole gamut of tasks; ranging from asset selection, portfolio allocation, risk management, regime detection, and so on. I have recently been working on these in-house engines as well as the software architecture and engineering at Ask2.ai.

Accumulation Point: ML, AI, statistics, and optimization consulting

I co-founded and work with Accumulation Point, a consulting firm that provides services in AI, LLMs, machine learning, statistics, and optimization to a wide variety of clients. My business partner in this venture is a long-time friend and colleague, Prof. Yoni Nazarathy based in Brisbane, Australia.

Fireball: using ML to rapidly detect and act on bush-fires

In the past half decade, I’ve worked with Fireball to build an ML-based system that spots wildfires from cameras placed in hilltops in the Sierra Mountains and elsewhere. We use custom ML-models to spot the fires, then triangulate the location of the base of the fire from cameras via raytracing, and alert local emergency services. I designed and architected the overall system and have worked on various improvements on it, ranging from ML research for better per-image accuracy, to implementing star-based camera and raytracing calibration, and then to building caching infrastructure to reduce time-to-alert. It’s been very fun and rewarding, and I’ve learned a whole lot about our physical world and how one should model it mathematically and in software!

Interests

I have a deep curiosity for understanding this world and what drives it forward, and I’m interested in doing things that have a significant positive impact on people’s lives. I’m basically interested in anything where I get to solve tough but meaningful problems with intelligent, passionate, and creative people.

I think technology is one of the most useful and powerful things we’ve come up with, so I’ve spent significant time becoming very good at wielding technology to get stuff done. I like the ethos of “hacking on things” (in the non-security way) and understanding things to the lowest bit-and-byte level, and this has made me good at throwing things together in order to make stuff happen. I can quickly build both small and large systems, and I hope that you can see a bit of that on this site and on my Blog. I mostly use the blog to collect things I feel might be useful to refer to later, but it also showcases the kinds of things I tinker with when not working on bigger projects.

I find cryptography very interesting, and I have extensive professional in both practical (SSL/TLS, X.509, PGP, HSM, primitives etc), and theoretical (symmetric, asymmetric, zero-knowledge, RSA, ECC, etc) cryptography.

I love travel, and in particular I enjoy visiting places that are on the somewhat less beaten path while seeing them from the eyes of locals. You can find out more about my travel experiences (as well as a cute map of everywhere I’ve been) on my Couchers profile.

I love geography, earth science, as well as maps and cartography. I find our physical planet and differences in different geographic locations affect the peoples living in them endlessly fascinating. My home tends to be covered in various maps. I try to contribute to (as well as make use of) OpenStreetMap whenever I get a chance.

I maintain a small Guide to NYC with places to eat and things to do! I’ve neglected it for some time though, so there’s not that many recently reviewed places.

I like building little websites (like this one) on the “vanilla web”. I try to stay away from big social media sites and I wish more people and projects would build their own sites out in the open instead of within those walled gardens.

Education

I studied undergraduate mathematics at the University of Queensland, in Brisbane, Australia. During my undergraduate years, I studied under the guidance of Yoni Nazarathy and Huy Nguyen to whom I’m indebted for their invaluable advice in life and in maths. I studied courses in mathematics, physics, statistics, and economics among others, and graduated at the top of my class.

I went on exchange to the University of California, Berkeley in the last year of my undergraduate degree where I studied graduate analysis, differential equations, and differential geometry.

After my undergraduate degree, I moved to Melbourne to complete a masters degree in statistics and mathematics at the University of Melbourne. My research advisor was Peter Taylor and I focused on studying the block propagation process of the Bitcoin network, and modelling its propagation characteristics. My thesis is available online and you can download the data and read more about what I did at this dedicated website.

I’m currently completing my PhD in Operations Research at Columbia University.