Donald Kong
theoretical cs student :)
NUS × Télécom Paris · complexity & cryptography
about
I'm a computer science undergraduate at NUS, spending 2026–28 in France on the NUS–Télécom Paris double degree. I like problems that look simple and aren't; right now that means computational complexity and cryptography: what makes problems hard, and how to put that hardness to work.
Alongside computer science I take a second major in Statistics, including Bayesian statistics and stochastic processes. I speak English, Mandarin and Cantonese natively, and French at B1 (DELF).
code Python · C++ · Java · TypeScript tools LaTeX · Git · Linux · Vim
timeline
now
Engineering years in France.
The first-year engineering core is taught in French. Coursework includes probability, functional and Fourier analysis, and compilers.
role
Diplôme d'Ingénieur student, Télécom Paris (Sep 2026 – Aug 2028)
topics
- Probability
- Functional analysis
- Compilers
links
Reproduced the official ML-KEM decryption-failure figures, then bounded the failure rate, at DSO National Laboratories.
ML-KEM (Kyber) is the NIST post-quantum KEM standard. The estimates and provable bounds came from Chernoff and saddlepoint methods, checked against exact computation. The work ended as a five-chapter research report and a 45-minute talk.
role
Researcher (Intern), DSO National Laboratories (May – Aug 2026)
topics
- ML-KEM (Kyber)
- Chernoff bounds
- Saddlepoint method
links
Studios for CS1101S, then tutorials for CS2040S.
CS1101S teaches functional programming through an SICP-style curriculum in Source; CS2040S added consultations, and I graded assessments for both. Explaining amortised analysis, or why a particular invariant holds, is a different skill from doing the analysis, and it turned out to be one I enjoy.
Design and analysis of algorithms (CS3230) and theory of computation (CS3231).
These were the modules that did it: complexity classes, reductions, what's hard and why. This was when "theoretical CS" stopped being a vague label and started being a concrete set of things I wanted to keep working on.
Matriculated at NUS: the year I worked out where my interest sits.
A+ in CS2040S (data structures and algorithms), A+ in CS1101S (programming methodology), A in CS1231S (discrete structures). Realised somewhere in the middle of CS1231S that the mathematical and theoretical side of CS is what I actually care about, more than software engineering. The rest of the degree has been shaped by that.
Singapore's mandatory military service. Off-keyboard, mostly.
Not much computer science happened here, but the reading list got long.
role
Full-time National Serviceman (Jan 2022 – Nov 2023)
Deep learning models for face recognition, face detection and emotion recognition.
I designed and evaluated the models, and helped improve them and deploy them in production computer vision pipelines. A lot of the time went into data collection, accuracy/speed trade-offs and integration with the rest of the system.
role
AI Scientist (Intern), Pensees Systems (Sep – Dec 2021)
topics
- Deep learning
- Computer vision
First proper research project, at SUTD while still in high school (NUS High).
Analysed how set similarity, cardinality and the number of comparison bits affect estimator bias in b-Bit Minwise Hashing, and proposed an error-term correction to the similarity estimator with negligible runtime and memory overhead.
role
Student Researcher, Singapore University of Technology and Design (Jun 2019 – Feb 2020)
topics
- Algorithmic analysis
- Estimator bias
- Set similarity
contact
up for a chat?
I'm looking for research internships in complexity or cryptography for summer 2027. Email works best; I'm on Central European time until 2028.