I’m a researcher and builder focused on applied AI, data visualization, and human-centric information systems. I love turning complex legal and policy workflows into interactive, trustworthy experiences. Outside of work I train powerlifting and practice judo — discipline and iteration carry over into my engineering.
- Based in the UK • collaborating with universities and research teams
- Obsessed with: knowledge graphs, retrieval-augmented generation (RAG), and visual analytics
- Currently shipping: PoliViz, Emploodr, and LexVizExplorer
Interactive legal-process explorer built with React + TypeScript. It blends semantic flow selection, graph validation, and embeddings-backed retrieval to guide users through complex administrative justice pathways.
Key ideas recently shipped:
- aj_* ID standardization across Administrative Justice templates for clean registries and semantic matching
- Hybrid semantic+keyword flow selection with graceful fallback when embeddings are stale
- Flow integrity validator and Jest suites (green!) to prevent regressions
- Embeddings pipeline with non-destructive updates to preserve curated descriptors
Transform complex privacy policies into clear, actionable insights. POLIVIZ detects policies on any site, runs GPT‑4 analysis with confidence/risk scoring, answers questions in real time, and extracts GDPR‑relevant information — all with privacy-first design.
- Visual argument mapping • provenance-aware snippets • literature-aware Q&A
- Designed for research labs and public policy teams
AI-enhanced tool guiding people through employment law using GPT‑3.5‑Turbo with RAG. Provides personalized ODR tool recommendations, side‑by‑side comparisons with relevance/cost/accessibility scoring, and a curated knowledge base for England & Wales with real‑time triage.
- Languages: TypeScript, Python, SQL
- Frontend: React, Next.js, D3, Cytoscape.js, XFlow/React Flow, Styled Components
- Testing/Quality: Jest, React Testing Library, Playwright, ESLint, Prettier
- Backend: Node.js/Express, FastAPI, REST APIs, WebSockets
- AI/ML: OpenAI APIs, RAG, embeddings, semantic search, prompt/eval pipelines
- DevOps/Infra: Docker, CI/CD, GitHub Actions, PostgreSQL, Vercel/Azure/AWS
- Visualization: graph layouts, timelines, dashboards, interaction design
- Human-centered information retrieval and evaluation for RAG systems
- Knowledge graphs + policy/legal process modeling
- Visual analytics for explainable AI and decision support
- Responsible AI: transparency, provenance, and bias mitigation
- NITDA Machine Learning Professional — Issued Dec 28, 2023
- IBM Data Science Professional Certificate — Issued Apr 10, 2023
- IBM Data Analyst Professional Certificate — Issued Apr 16, 2023
- Applied Data Science Specialization — Issued Apr 16, 2023
- Data Visualization & Dashboard Essentials — Issued Apr 4, 2023
- Venue: IEEE VIS 2025 — The premier forum for visualization and visual analytics (Jul 23, 2025)
- Summary: Presents EmployODR, an AI-enhanced online tool for employment-law dispute guidance. Integrates GPT‑3.5 Turbo with structured legal data via a RAG pipeline; uses node‑link network maps and plain‑language summaries to communicate reasoning and guide users through complex scenarios. Explores how visualization principles can promote transparency in AI-driven legal decision-making and improve access to justice.
- Venue: Law Society of England and Wales (Jun 23, 2025)
- Summary: EmployODR featured in the final report for its contribution to publicly funded online legal guidance. Highlights AI-powered legal guidance, innovation to bridge the justice gap, and publicly accessible tools as part of modernizing access to justice in England and Wales.
- Venue: ICAIL 2025 — AI for Access to Justice Workshop (Jun 20, 2025)
- Summary: Describes the EmployODR prototype developed under the UK Regulators Pioneer Fund 3. Assesses the use of LLMs for recommending appropriate ODR tools in employment law, covering development insights and user testing to inform future research and deployment.
Crafting Visual Narratives: A Case Study on Developing an Engaging Visualisation Poster Using U.S. Immigration Data
- Venue: Eurographics Digital Library (Aug 31, 2023)
- Summary: A design study outlining a methodical process for turning raw immigration data into a compelling poster narrative. Emphasizes a “big hero” visualization, coherent narrative flow, and well-curated meta information to make the poster self-explanatory.
Interweaving Data and Stories: Unveiling the Human Dimension of U.S. Refugee Movements through Narrative Visualisation
- Venue: Eurographics Digital Library (Aug 31, 2023)
- Summary: An interactive narrative visualization using Mapbox and D3.js to connect geographic and temporal aspects of U.S. refugee data. Demonstrates modular preprocessing and interaction techniques to foster empathetic understanding of refugee dynamics.
- Venue: Eurographics Digital Library (Aug 31, 2023)
- Summary: A WebXR tool built with A‑Frame, D3.js, and JavaScript for immersive exploration of complex datasets. Proposes a user study using NASA TLX to assess usability and workload; discusses future directions for guidance, layout optimization, and long-term UX.
- Support international applicants with admissions, CAS, and visa processes; provide 1:1 guidance via phone/email; collaborate across teams to ensure a smooth journey.
- Lead tours, represent the university at events, and advise prospective students; contribute feedback to improve student engagement.
- Demonstrator for ICE‑3703/4702 (Machine Learning/Advanced ML) in the School of Computer Science & Engineering.
Artificial Intelligence Researcher — Solicitors Regulation Authority (May 2024 – Feb 2025, Contract)
- Built an AI‑assisted legal visualization tool for employment‑law needs assessment and ODR recommendations.
- Lead engineering for product features and testing; drive end‑to‑end SDLC quality.
- Co‑founded and shipped solutions for SMEs; designed test strategies, led development across full SDLC, and delivered data‑driven features.
- Focus: AI/ML, visualization for legal advice systems; activities: Course Rep Trainer, Peer Guide, International Student Mentor.
- Grade: Distinction; Supervisor: Henning Flaecher; Programme: 2PHYS028R.
- LexVizExplorer: Administrative Justice flow revamp (aj_* IDs), validator/test automation, semantic selection with robust fallbacks
- PoliViz: Research timeline mining + debate-graph visualization + citation-aware retrieval
- Emploodr: Skills/competency graph modeling and explainable matching prototypes
- Powerlifting (strength and discipline) • Judo (leverage and control) • Always iterating
- 🥋 Judo and 🏋️ Powerlifting keep me sharp — balance, leverage, and patience carry into my engineering flow.
- 🎨 I love making data “feel” interactive — tiny details, smooth motion, and just the right contrast.
- 🌱 Always learning: better evals for RAG, smarter graph layouts, and ways to demystify model behavior.