MEng Computer Science at UCL
About Me
Syed Ali Abbas
MEng Computer Science at UCL
Computer Science | AI Systems | Robotics | Applied ML
I am a Computer Science MEng student at University College London with a First Class average and a strong focus on AI systems engineering, applied machine learning, and robotics. I build end-to-end, production-ready platforms that combine rigorous engineering, modern AI techniques, and thoughtful user experience from Retrieval-Augmented Generation systems to global EdTech platforms.
2024-2026
Featured Projects
Flagship projects spanning AI systems, robotics, and full-stack development.
Hatchmark
Blockchain-powered content authenticity platform
A blockchain-based platform for verifying content authenticity using AI detection, built for ETHOxford 2026. Implemented with Sui Move smart contracts, @mysten/dapp-kit for Web3 integration, Supabase for backend services, and a Chrome Extension for real-time content verification. Combines on-chain provenance tracking with AI-powered deepfake and manipulation detection to establish digital content authenticity.
Enermation
Automotive export/import marketplace with AI chatbot
A full-stack automotive marketplace for vehicle export and import operations, featuring an AI-powered chatbot for customer queries, integrated with Shopify Storefront API for product management and Stripe for payment processing. Uses Groq AI for fast LLM inference in the chatbot, Google Vision API for vehicle image analysis and condition assessment, and Next.js 16 for a performant frontend with real-time inventory tracking.
Poligent
AI Policy Simulation Platform — multi-agent fiscal policy analysis
A multi-agent AI system generating structured stakeholder reasoning for proposed fiscal policies. Built during the Cerebral Valley x Anthropic "Built with Opus 4.7" hackathon — selected from 13,000+ applicants. Features a supervisor agent + 4 demographically-grounded archetype agents (ONS/IFS-sourced personas: Sarah the single parent, Mark the sole trader, Priya the financial analyst, Arthur the retired worker) running in parallel with real-time streaming. Synthesises first-person reactions into a structured policy brief with distributional impact analysis, validated against IFS published distributional findings. React dashboard + Streamlit fallback + CLI. Tech: TypeScript, Python, React, Docker.
Sapiens Nova Academy
End-to-end EdTech platform with AI-powered learning assistant
Sapiens Nova Academy is a global education and innovation platform designed to prepare students for careers in technology and entrepreneurship. Built with Next.js, TypeScript, Tailwind CSS, featuring a Retrieval-Augmented Generation (RAG) assistant powered by LangChain, integrated Stripe payment processing, and PostgreSQL database with pgvector for semantic search. The platform supports multi-programme education delivery, secure enrollment management, and real-time payment tracking.
Systems Engineering Intern at HP
Production RAG pipeline for enterprise document processing
Designed and implemented a Retrieval-Augmented Generation pipeline for enterprise document understanding at Hewlett-Packard. Built with Python, LangChain, Llama 3.1 8B (via Ollama), and Hugging Face embeddings. Deployed containerized on HP edge workstation using Docker. Features document ingestion, intelligent chunking, semantic retrieval using ChromaDB, and response generation workflows with comprehensive evaluation frameworks.
Real Estate Platform — Other Dev
High-performance property showcase platform with lead generation
Developed a production-ready real estate platform designed to showcase premium properties and drive client engagement. Built with Astro, React, TypeScript, and Tailwind CSS, achieving sub-2-second load times through efficient rendering strategies and careful asset management. Implemented dynamic property filtering, GSAP-based animations, and intuitive navigation flows tailored to real estate browsing behaviour. Integrated lead-generation workflows to convert user interest into actionable business enquiries, balancing technical performance with user experience and real business requirements.
ML Computer Vision Systems
CNN-based image processing and neural texture synthesis
Implemented advanced machine learning systems for visual computing, including CNN-based image denoising using PyTorch, neural texture synthesis using Gram matrices, and object detection and segmentation pipelines. Applied advanced image processing techniques in MATLAB and PyTorch, combining theoretical understanding of convolutional architectures with practical implementation for real-world visual tasks.
Professional Experience
Work Experience
Building production systems across AI engineering, fintech, and enterprise software.
T3 Consultants
- •Built VerifAI — an AI vendor discovery platform benchmarking 1,500+ AI vendors across accuracy, latency, cost, and integration complexity using semantic vector search and hybrid BM25 retrieval with Reciprocal Rank Fusion
- •Designed a 6-layer system architecture with Django Ninja async backend, Nuxt 4 frontend, and LangGraph-powered conversational discovery agent for enterprise procurement
- •Implemented evaluation pipelines using ROUGE/BLEU/BERTScore metrics with GDPR and EU AI Act compliance frameworks for responsible AI benchmarking
- •Developed taxonomy enrichment and auto-linking scripts using FastEmbed/Bedrock embeddings with BM25 for vendor-feature matching across 1,500+ vendors
Allica Bank
- •Automation Analyst Intern at a UK fintech unicorn (£1.2B valuation) specialising in commercial lending with Agentic AI and ML as core focus areas
- •Part of UCL Year in Industry programme, working in the Lending department
Hewlett-Packard
- •Designed and implemented a production-grade RAG pipeline for enterprise document understanding using Llama 3.1 8B, LangChain, and Hugging Face embeddings
- •Containerized deployment with Docker on Zbyhp Edge workstation for reproducible local development and enterprise scaling
- •Built ChromaDB-backed vector store for semantic retrieval with intelligent document chunking strategies
- •Developed comprehensive evaluation frameworks assessing model accuracy, response grounding, and system reliability
Technical Proficiency
Skills
Technologies and tools I work with across AI, full-stack, and embedded systems.
Programming Languages
AI & Machine Learning
Frameworks & Tools
Robotics & Embedded
Beyond the Classroom
Leadership & Achievements
Recognised for leadership, event scale, and impact across education and community initiatives.
Head of Robotics & Physics
SCINNOVA Science Olympiad
- •Led a team of 10 to design and deliver robotics and physics modules for SCINNOVA VI — a national science olympiad
- •Awarded for Excellence in event planning and leadership, recognised for exceptional organisational impact
- •Designed curriculum for robotics competitions and physics problem-solving tracks
Executive Member
Cedar College Robotics & Engineering Society
- •Organised robotics and web development workshops for over 100 students, teaching circuit design and programming fundamentals
- •Led robotics projects including self-balancing robots, line-following systems, obstacle-avoiding robots, and Bluetooth-controlled vehicles
- •Installed automated campus facilities including smart lighting systems and automated water management — reducing water wastage by ~30%
Get In Touch
Let's Build Something Impactful
I'm actively seeking internship opportunities, industrial placements, and research collaborations in AI systems, applied machine learning, and robotics. Drop me a message and let's discuss how we can work together.