All projects

Towards Understanding LLM Generalization through Fine-tuning

SPAR · ICML 2026 workshop paper

How finetuning on a small, narrow set of factual documents affects the actions a model takes on downstream tasks. First author on the paper, which was presented at an ICML 2026 workshop. Done as part of SPAR with mentors now at Redwood Research (Keshav Shenoy) and Anthropic (Emil Ryd).

paper code

Agent swarm evaluation of Android cyber capabilities

IMDA · Singapore AI Safety Institute*

Designed and built a swarm-based Inspect evaluation of frontier models' cyber capabilities against Android targets. Worked in IMDA's Biz-Tech group, which works closely with the Singapore AI Safety Institute. Also presented key takeaways from the HuggingFace–OpenAI and other loss-of-control incidents to decision makers.

code testimonial

Cambridge AI Safety Hub

Committee member · Alignment Fellowship · Alignment Desk

Currently on the student committee. Previously completed the six-week Alignment Fellowship reading group and wrote two posts as part of the Alignment Desk writing group.

writeup 1 writeup 2

Alignment Research Bootcamp Oxford (ARBOx)

Oxford

Two-week technical bootcamp working through the ARENA curriculum. Work included building GPT-2 from scratch, reinforcement learning and mechanistic interpretability.

Image embeddings from small datasets

DSO National Laboratories · Research intern

Researched and implemented a method for generating image embeddings from small datasets, building a documented Python codebase from scratch.

code testimonial

Dye-sensitised solar cells

National Institute of Education · Research intern

Investigated dye-sensitised solar cells with Prof. Tsao Hoi Nok. Delivered a poster and contributed to a published research paper.

paper

Aerial vehicle classification

DSTA · Research intern

Built a computer vision model to detect and classify vehicles in aerial imagery; one of four projects selected to present on stage at the 2019 NUS High Research Congress.

Integrated Design Project robot — first place

Cambridge Engineering · Project and software lead

Built a competition robot from scratch in a six-person team and took first place across the year group. Led software and project management across the hardware, electrical and software sub-teams. Features include a simulator, test suite and visualiser.

code

Construction robot mapping system

Fabrica AI · Software engineering intern

Integrated a new LiDAR into a construction robot's mapping system in a Dockerised ROS 2 environment. Also created a schema to verify robot configuration changes. Worked in Python and C++.

testimonial

Drone countermeasure systems

DSTA · Engineering intern

Developed a drone used for drone countermeasures. Upgraded the video transmission scheme, raising frame rate by 30% and cutting bandwidth by 75%. Carried out a network architecture migration and improved the ground control application. Worked in Java and Python.

testimonial

F-16 aircraft technician

Republic of Singapore Air Force · National service

Serviced F-16s for daily training flights. Chosen as the Operational Airman of the Month (June 2023). Selected for several overseas deployments.

Ranked 30th of 320 (Year 2)

Cambridge Engineering · MEng Information Engineering

transcript

Ranked 14th of 323 (Year 1)

Cambridge Engineering · MEng Information Engineering

transcript

Operational Airman of the Month

Republic of Singapore Air Force

NUS High School Diploma, 4.9/5.0 GPA

Principal's List (top 5%)

Silver, Asian Physics Olympiad

33rd globally · 2nd in Singapore

website

Gold, Singapore Physics Olympiad

Top 28 nationally

website

Gold, Singapore Junior Physics and Chemistry Olympiads

Gold in both olympiads, two years running