CV
Download CV (PDF)
Machine learning researcher working across computer vision, 3D human modelling, differentiable radar simulation, foundation-model transfer, and sensor-based perception. I build end-to-end AI systems — from dataset collection and model design through rigorous evaluation to deployment on real sensing hardware — with prior industry experience leading production software teams.
Education
- PhD in Artificial Intelligence, University of Southampton, 2022–present
- UKRI MINDS CDT Studentship (supervisor: Dr Shelly Vishwakarma)
- MSc in Artificial Intelligence (Distinction), University of Southampton, 2021–2022
- Thesis: Universal Gene Transformer: Cancer Subtype Classification using RNA-Seq Data (supervisor: Dr Jo Grundy)
- BSc in Computer Science (Distinction), University of the West Indies, Mona, Jamaica, 2014–2017
- Thesis: An Evolutionary Approach to Training Intelligent Agents using NEAT (supervisor: Dr Gunjan Mansingh)
Professional experience
PhD Researcher — Radar-based AI, Human Motion Modelling & Sensor Simulation University of Southampton · 2022–present
- Developed a temporally coherent whole-body 3D human mesh recovery framework for reconstructing motion from monocular video, using PyTorch, LoRA/PEFT fine-tuning, mixed-precision training, and model distillation; evaluated across 12 datasets (~12M samples) with state-of-the-art whole-body accuracy (manuscript under review)
- Built a differentiable sensor-simulation pipeline that generates synthetic radar micro-Doppler data directly from 3D human motion, enabling scalable training-data augmentation for activity and gesture recognition without new radar captures (manuscript under review)
- Adapted vision foundation models (DINOv2, CLIP, MAE) for radar and WiFi spectrogram recognition, studying cross-modal transfer under sensor shift (manuscript under review)
- Designed multimodal data collection and evaluation pipelines integrating TI mmWave radar, Azure Kinect, OpenCV, MATLAB, and signal-processing workflows — including radar-based neuroimaging work published at IEEE EMBC 2024 (oral presentation; NextGen Scholar Award)
- Optimised ML workflows for efficient training and deployment using CUDA, TensorRT, NVIDIA DALI, mixed precision, and model compression
- Developed robotics and edge-AI prototypes with ROS, NVIDIA Jetson, Arduino, and MQTT for heterogeneous swarm robotics and disaster-response applications
Masters Researcher University of Southampton · 2021–2022
- Built an end-to-end deep learning model combining 1D convolutions and multi-head self-attention for cancer subtype classification from high-dimensional gene expression data, achieving 92–94% accuracy across 33 subtypes and outperforming deep learning and classical baselines
- Designed an explainability pipeline using saliency maps and gene-effect scores to identify candidate biomarkers, validated against established cancer databases
- Reimplemented a self-supervised motion-transfer method for animating still images via latent-space transformations from driving video
Technical Lead / Scrum Master / Lead Software Developer NCB Financial Group, Kingston, Jamaica · 2018–2021
- Led delivery of Jamaica’s first online mortgage portal and centralised lending platform, reducing application time from weeks to 3 days and digitising 80% of the lending process
- Architected microservices, testing, and analytics infrastructure; contributed to $200M+ annual cost reduction and drove Kubernetes enablement across engineering teams
- Presented architecture proposals and delivery updates to C-suite stakeholders
Technical skills
- AI & deep learning: PyTorch, TensorFlow, Hugging Face Transformers, DINOv2, CLIP, MAE, DeepSpeed, RAPIDS cuML, TensorRT
- Training & optimisation: LoRA, mixed precision, model distillation, multi-GPU / multi-node training, HPC, DALI
- Architectures: Transformers, 1D CNNs, Mamba, Mixture of Experts, Diffusion, Flow Matching
- LLM & retrieval: RAG pipelines, embedding models, reranking, vector search, agentic workflows
- Evaluation: benchmark design, ablation studies, cross-domain evaluation, macro-F1, MCC, cross-validation, retrieval metrics
- Programming: Python, C++, Java, JavaScript, TypeScript, MATLAB
- 3D & simulation: SMPL-X, differentiable simulation / rendering, Ansys HFSS, Blender, Open3D, Trimesh, PyVista, OpenCV
- Hardware & sensors: FMCW radar, micro-Doppler, TI 77 GHz radar, Infineon radar, Azure Kinect / RGB-D, Jetson, Raspberry Pi, ROS
- Infrastructure: Docker, Kubernetes, CI/CD, Git
Publications
K. Peart, I. Bodala and S. Vishwakarma, "Advancements in Non-Invasive Neuroimaging: Exploring the Potential of Radar Technology for Brain Imaging and Tumour Detection," 2024 46th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Orlando, FL, USA, 2024, pp. 1-6, doi: 10.1109/EMBC53108.2024.10782724.
Three further first-author manuscripts are currently under review at international machine learning and computer vision venues.
Talks & demonstrations
2026
Live demonstration at ECS Student Showcase 2026
2025
Symposium talk at MINDS CDT Symposium, University of Southampton, Southampton, UK
2025
Public engagement demonstration at Southampton Science and Engineering Festival (SOTSEF), Southampton, UK
July 2024
Conference oral presentation at IEEE EMBC 2024, Orlando, Florida, USA
July 2024
Symposium talk at MobiUK 2024 — Sixth UK Mobile, Wearable and Ubiquitous Systems Research Symposium, Southampton, UK
Teaching
Awards & scholarships
- IEEE EMBC NextGen Scholar Award (2024)
- Best Demonstration Award, ECS Student Showcase (2026)
- Chevening Scholarship (2021)
- UKRI MINDS CDT Studentship (2022)
- Dean’s List (2016–2017), Letters of Commendation (2014–2016), UWIDEF and Flow Foundation Jamaica Merit-Based Scholarships — University of the West Indies
Certifications
- Neural Networks and Deep Learning — DeepLearning.AI
- AI For Everyone — DeepLearning.AI
- Advanced Certified Scrum Developer (A-CSD) — Scrum Alliance
- Certified ScrumMaster (CSM) — Scrum Alliance
- Agile Project Management (ICP-ACM) — ICAgile / Cprime
- Enterprise Design Thinking Practitioner & Team Essentials for AI — IBM