Hi, I’m Keniel 👋
I’m a PhD researcher in Artificial Intelligence at the University of Southampton, part of the UKRI MINDS Centre for Doctoral Training. I work at the intersection of computer vision, 3D human motion modelling, and radar sensing — building AI that understands how people move, even when a camera can’t see them.
The thread running through my research is human sensing, end to end: I build the datasets, the models, the simulators, and the evaluation pipelines — and I care that results hold up on real hardware, not just on benchmarks.
- Computer Vision
- Human Pose Estimation
- Radar & Wireless Sensing
- Multimodal Learning
- Foundation Models
- Simulation & Synthetic Data
News
For DJ Air, a real-time gesture-controlled DJ system running entirely in the browser.
Spoke on expressive human pose estimation at the MINDS CDT Symposium; demoed the AI DJ at the Southampton Science and Engineering Festival.
Oral presentation of our radar neuroimaging work at IEEE EMBC 2024 in Orlando; also presented at MobiUK 2024.
Funded by a UKRI MINDS Centre for Doctoral Training studentship.
Came to the UK as a Chevening Scholar for the MSc in AI at Southampton, completed with Distinction.
What I work on
🧍 3D Humans from Video
Temporally coherent whole-body mesh recovery — body, hands, and face together — from ordinary monocular video, built on large vision transformers and evaluated across 12 datasets (~12M samples).
📡 Differentiable Radar Simulation
Rendering physically faithful radar micro-Doppler straight from animated 3D human meshes, so synthetic data can stand in for expensive radar captures.
🧠 Foundation Models Beyond Vision
How far do frozen DINOv2, CLIP, and MAE features transfer to radar and WiFi spectrograms? Further than you'd think — it's a representation-transfer problem.
🩺 Radar Neuroimaging
Feasibility of radar for non-invasive brain imaging and tumour detection — published at IEEE EMBC 2024 (oral presentation, NextGen Scholar Award).
Several first-author manuscripts from this work are currently under review at international venues — the research page has the high-level story of each.
Education
2022 – present · UKRI MINDS CDT Studentship · computer vision, 3D human motion & radar sensing (supervisor: Dr Shelly Vishwakarma).
2021 – 2022 · Chevening Scholar · thesis on cancer subtype classification from RNA-seq data (supervisor: Dr Jo Grundy).
2014 – 2017 · Dean's List · thesis on training intelligent agents with neuroevolution (supervisor: Dr Gunjan Mansingh).
Before the PhD
I spent three years at NCB Financial Group — the Caribbean’s largest financial group — as a technical lead, Scrum master, and lead software developer. I led delivery of Jamaica’s first online mortgage portal and a centralised lending platform, digitising 80% of the lending process and contributing to $200M+ in annual cost reductions. That production-engineering background still shapes how I do research: reproducible pipelines, real deployments, and systems other people can build on. Originally from Kingston, Jamaica 🇯🇲.
Get in touch
The fastest way to reach me is email at peartkeniel@gmail.com. I’m also on GitHub, LinkedIn, Google Scholar, and ORCID.
