I am a PhD student in Engineering Mathematics at the University of Bristol, working around tactile robotics, neuromorphic touch, and embodied intelligence.
Recently I have been building open-source agent skills for self-learning: tools that help beginners turn "I want to learn X" into a path, a project, a check question, and a feedback loop.
| Area | What I am exploring |
|---|---|
| Embodied AI and robotics | perception-action loops, tactile sensing, simulation-first learning, robot policy debugging |
| Neuromorphic touch | event-based tactile sensing, texture, contact, force, speed, and human touch behavior |
| Agent-assisted learning | skill packs, AI tutors, paper-to-practice workflows, codebase-to-course learning |
| Research tooling | literature maps, reproduction plans, lab workflows, verification, and source trust |
| Project | Why it matters |
|---|---|
| learn-anything-skills | 100 beginner-first agent skills for learning technical fields through maps, projects, tests, and reflection. |
| embodied-ai-robotics-skills | A focused skill pack for learning embodied AI through simulation, perception-action loops, and robot policy debugging. |
| ai-paper-reading-skills | Skills for turning AI papers into task setups, method cards, reproduction plans, and citation maps. |
| codebase-apprentice-skills | Skills for turning real repositories into guided beginner courses and first contribution paths. |
| verification-trust-skills | Skills for checking claims, auditing sources, detecting hallucinations, and labeling uncertainty. |
| Human-Texture-Exploration-Experiment | Research code around human touch, force, speed, and texture exploration. |
Full satellite skill index: learn-anything-skills/docs/SATELLITE_REPOS.md
learn-anything-skills
-> AI paper reading
-> embodied AI and robotics
-> codebase apprenticeship
-> human writing
-> verification and trust
-> learning OS productivity
-> startup, market, career, API, data, and research workflows
The design rule is simple:
next action + visible artifact + check question + what to ignore + reflection
- tactile robotics and neuromorphic touch
- embodied intelligence and robot learning
- human texture exploration and contact behavior
- AI-assisted research workflows
- beginner-friendly technical education
Python · Jupyter · NumPy · pandas · scikit-learn · matplotlib · GitHub Actions · Markdown · LaTeX
I like building small systems that make hard fields less lonely to enter. My current open-source direction is practical: if a tool does not help someone decide what to do today, it is not finished yet.