09/30/2026 updated
100 % available
Remote Full-Stack AI Engineer | Research, Development & Deployment
Manchester, United Kingdom
Only remote
MSc Dependable Software Systems (Distinction), St Andrews & Maynooth; BSc Computer Science, NUSTAbout me
Full-Stack AI Engineer, about 10 years across THG, Gifftid and research. I build AI products end to end: RAG, tool-using agents, LLM guardrails, Next.js + Cloud Run. UK sole trader based in Tokyo, working remotely with clients in any time zone.
DevOpsTensorFlowAzure Machine LearningNext.jsRust (Programming Language)Prompt EngineeringGenerative AIFull Stack DevelopmentMachine Learning Operations
AI & Machine Learning Engineering
Expertise in building production-ready AI systems including RAG pipelines, LLM grounding, tool-using agents, agentic AI, fine-tuning, prompt engineering, multi-modal models, NLP, computer vision, and classical ML. Hands-on experience with GPT-4/5, Claude, Gemini, Llama, BERT, RoBERTa, DistilBERT, BLIP-2, InstructBLIP, and CLIP models deployed at scale serving 7M+ monthly users.
Full-Stack Development (Next.js, React, FastAPI, Python, TypeScript)
End-to-end development of scalable web applications and APIs using Next.js, React, Node.js, FastAPI, Python, TypeScript, JavaScript, Tailwind CSS, Flask, and Tauri. Delivery of complete SaaS platforms, public REST APIs, and native desktop applications built and shipped solo.
Cloud & DevOps (GCP, Firebase, Docker, Kubernetes, CI/CD)
Architecture and deployment of cloud-native solutions on GCP (Cloud Run, BigQuery, Vertex AI, Cloud Functions, Pub/Sub, IAM), Firebase, Firestore, and Azure. Containerised deployments with Docker and Kubernetes, CI/CD pipelines via GitHub Actions, and serverless microservices architectures.
LLM Tooling & Orchestration
Proficiency with OpenAI, Anthropic Claude, Anthropic SDK, Claude Code, Gemini Function Calling, LangChain, LangGraph, LlamaIndex, NeMo-Guardrails, DialogFlow, BigQuery Vector Search, Pinecone, Chroma, ONNX, and Ollama for building and orchestrating LLM-powered applications.
Data Science & Analytics
Topic detection and sentiment analysis using guided LDA, RoBERTa, BERT, and DistilBERT; data pipelines on BigQuery, PostgreSQL, MySQL, and ETL workflows; analytics dashboards with Recharts, Plotly, and Streamlit; time-series forecasting, recommendation systems, and A/B testing.
ML Frameworks & Libraries
Hands-on experience with PyTorch, TensorFlow, Keras, scikit-learn, Transformers, Hugging Face, spaCy, NLTK, Gensim, Pandas, and NumPy for research, prototyping, and production ML pipelines.
Red-Teaming & Model Evaluation
Red-teaming LLM outputs adversarially, running model-response evaluation passes on major LLM features, implementing hallucination prevention and moderation guardrails, and conducting LLM evaluation including retrieval measurement (MRR) on golden question sets.
Agentic AI Systems Design
Design and solo build of tool-using LLM agents including customer-support agents, agentic RAG systems with hybrid search, and browser-only audit agents. Agents enforce money rules in tool code, plan their own searches, cite exact source sentences, and drive both Claude and Gemini from a single tool schema.
Rust & Native Desktop Development
Signed, notarised desktop software built and shipped solo using Rust and Tauri, including a privacy browser, native IDE, GPU photo editor, video player, terminal emulator, and timezone calendar, published at ibraryunus.com/ai-engineer.
Expertise in building production-ready AI systems including RAG pipelines, LLM grounding, tool-using agents, agentic AI, fine-tuning, prompt engineering, multi-modal models, NLP, computer vision, and classical ML. Hands-on experience with GPT-4/5, Claude, Gemini, Llama, BERT, RoBERTa, DistilBERT, BLIP-2, InstructBLIP, and CLIP models deployed at scale serving 7M+ monthly users.
Full-Stack Development (Next.js, React, FastAPI, Python, TypeScript)
End-to-end development of scalable web applications and APIs using Next.js, React, Node.js, FastAPI, Python, TypeScript, JavaScript, Tailwind CSS, Flask, and Tauri. Delivery of complete SaaS platforms, public REST APIs, and native desktop applications built and shipped solo.
Cloud & DevOps (GCP, Firebase, Docker, Kubernetes, CI/CD)
Architecture and deployment of cloud-native solutions on GCP (Cloud Run, BigQuery, Vertex AI, Cloud Functions, Pub/Sub, IAM), Firebase, Firestore, and Azure. Containerised deployments with Docker and Kubernetes, CI/CD pipelines via GitHub Actions, and serverless microservices architectures.
LLM Tooling & Orchestration
Proficiency with OpenAI, Anthropic Claude, Anthropic SDK, Claude Code, Gemini Function Calling, LangChain, LangGraph, LlamaIndex, NeMo-Guardrails, DialogFlow, BigQuery Vector Search, Pinecone, Chroma, ONNX, and Ollama for building and orchestrating LLM-powered applications.
Data Science & Analytics
Topic detection and sentiment analysis using guided LDA, RoBERTa, BERT, and DistilBERT; data pipelines on BigQuery, PostgreSQL, MySQL, and ETL workflows; analytics dashboards with Recharts, Plotly, and Streamlit; time-series forecasting, recommendation systems, and A/B testing.
ML Frameworks & Libraries
Hands-on experience with PyTorch, TensorFlow, Keras, scikit-learn, Transformers, Hugging Face, spaCy, NLTK, Gensim, Pandas, and NumPy for research, prototyping, and production ML pipelines.
Red-Teaming & Model Evaluation
Red-teaming LLM outputs adversarially, running model-response evaluation passes on major LLM features, implementing hallucination prevention and moderation guardrails, and conducting LLM evaluation including retrieval measurement (MRR) on golden question sets.
Agentic AI Systems Design
Design and solo build of tool-using LLM agents including customer-support agents, agentic RAG systems with hybrid search, and browser-only audit agents. Agents enforce money rules in tool code, plan their own searches, cite exact source sentences, and drive both Claude and Gemini from a single tool schema.
Rust & Native Desktop Development
Signed, notarised desktop software built and shipped solo using Rust and Tauri, including a privacy browser, native IDE, GPU photo editor, video player, terminal emulator, and timezone calendar, published at ibraryunus.com/ai-engineer.
Experience
Professional experience10+ years
Freelancer since2025
Languages
EnglishNative speakerThaiBasic knowledge
Project history
Sole trader serving multiple overseas clients remotely. Projects include: Gifftid R3+R4 SME Intelligence Platform (scoring UK SMEs on Risk, Readiness and Resilience using Next.js 16, TypeScript, Cloud Run, BigQuery, in-house AI systems, Stripe billing, Playwright and Vitest); Loredana AI Coaching SaaS (full-stack Next.js + Firebase platform with Gemini sentiment-aware coaching feedback and real-time Recharts dashboards); Telleosis/MyJamb EdTech AI Tutor (FastAPI backend on Cloud Run with RAG pipeline over BigQuery Vector Search, SQL gamification, leaderboards); Kotchakorn Language-Learning AI (interactive Thai-learning app in Python Shiny + BigQuery with Gemini 2.5 Flash streaming).
Architected and shipped the complete GenAI platform on GCP including CI/CD for Machine Learning, containerised deployments (Docker/Kubernetes), monitoring, access controls, chatbots, vector databases, and analytics dashboards. Ran red-team and model-response evaluation passes on each major LLM feature. Partnered with the CEO on strategic AI direction and shipped SME-focused tooling positioning Gifftid as a leading AI-driven SME platform.
Built production GPT chatbots grounded to company data with hallucination prevention, moderation guardrails, RAG, and chain-of-thought/tree-of-thought prompting (LangChain, NVIDIA NeMo Guardrails). Shipped multi-modal vision-language outfit recommendation chatbots using InstructBLIP and GPT models on Azure; deployed AI projects serving 7M+ monthly users. Built topic detection and sentiment-analysis pipelines using guided LDA, RoBERTa, BERT, and DistilBERT. Featured at THG Ingenuity Future of Commerce 2023 Industry Innovators Conference.