Computer Science × Mathematics Illinois Tech

Hello, I’m Belu.

Building toward robots that learn from experience.

I’m a researcher and builder interested in memory, machine learning, and intelligence in the physical world.

See what I’m exploring Say hello
01

Current direction

An open research question

Helping robots remember what matters.

I’m exploring memory systems for robots: how an embodied agent can carry useful experience across episodes without holding on to everything forever.

  1. 01

    Experience

  2. 02

    Remember

  3. 03

    Revise

  4. 04

    Act again

A loop, not an archive.

Today, a robot can finish one task and begin the next with very little of the judgment it just earned. I’m interested in the layer between perception and action that decides what becomes memory, what should be revised, and what is safe to forget.

The goal is simple to state and difficult to build: robots that become more capable through experience while staying adaptable when the world changes.

  1. 01

    What is worth remembering?

  2. 02

    When has a memory gone stale?

  3. 03

    How should one episode inform the next?

02

Selected work

Questions I’ve
learned by building.

Research and robotics projects that have shaped how I think about intelligent systems.

01

Generative models · Research

CrystalDiff-Q

Property-conditioned diffusion and transformer models for accelerating the search for stable quantum materials.

An exploration of generative modeling for crystal structures, connecting material properties with candidate generation.

  • Diffusion
  • Transformers
  • Materials
Read the paper View code
02

Computer vision · Research

Learning across skin tones

Investigating GAN-based image augmentation for more representative skin-disease classification.

This early project grew from a personal interest in diagnostic tools and the gaps created by underrepresented training data.

  • GANs
  • Computer vision
  • Health
Read the research View code
03

Robotics · Competition

FRC Team 2813

Programming competition robots to perceive, move, and complete tasks under real-world constraints.

With the Gear Heads, I helped build the software behind a regional-finalist season and a trip to the FIRST Championship.

  • Robotics
  • Controls
  • Team 2813
Meet the team
FRC Team 2813 with fellow competitors and their robots.
Field note / 01

FRC made robotics tangible: every clean idea eventually had to survive a noisy field, a hard deadline, and a real machine.

03

About

Lagos → Chicago

Curiosity, with somewhere to go.

I’m a sophomore studying Computer Science and Mathematics at Illinois Tech. I grew up in Lagos and moved to the United States at fourteen, beginning college a year later on a full scholarship.

My interest in machine learning began with a personal question. I lost my grandmother to an illness that went undiagnosed, and I wanted to understand whether better tools could help another family see what had been missed. That question led me first to medical imaging, then to generative models, and eventually to intelligence that can act in the world.

In 2026, I spent a summer at Apple working on computer vision for robots in recycling operations. More than a destination, it was a chance to learn from people building careful systems at scale. I’m carrying those lessons into the work I want to do next.

Outside research, I like the urgency of a hackathon, the elegance of an olympiad problem, and any excuse to make a machine move.

  • Apple, 2026

    Computer vision and robotics for e-waste recycling operations.

  • Hackathons

    Prototypes built at YC, Google, and South Park Commons events—including a robotic wheelchair.

  • Before ML

    Math and physics olympiads, plus a Rubik’s cube best of 18 seconds.
04

Contact

Start a conversation

Have a hard problem in the physical world?

I’d like to meet researchers and builders thinking seriously about robot memory, embodied learning, and useful AI.