Software & AI
Custom software, applied AI/ML, data and document automation, chatbots and integrations — scoped to the problem, not to a technology.
Most recent projects
See all projectsClaim document extraction for an insurance platform
Key-information extraction pulled VIN, odometer and ANR straight from photos.
Claim status chatbot for an insurance platform
Verified across four user roles and eleven real claims in structured acceptance testing, giving customers a plain-language status answer on demand instead of a support queue.
Aerial imagery pipeline for offline field mapping
Scattered public imagery, elevation and vector sources become one offline map package a field mapping app reads directly, rebuilt repeatably from a single command rather than assembled by hand.
Situational-awareness app for remote site operations
A role-simplified build of a specialist field coordination platform that ordinary crew can actually use, with its native dependency chain reproducible from a single containerised build instead of a hand-configured developer machine.
What I build
- Custom software
- Software built for the workflow you actually have, not a generic off-the-shelf template.
- Applied AI/ML
- Machine learning and AI applied to a real business problem — not a proof-of-concept that never ships.
- Data & document automation
- Extraction, classification and pipelines that turn messy documents and data into structured, usable output.
- Chatbots & assistants
- Conversational tools scoped to a specific task, not a general-purpose chatbot bolted onto a website.
- Integrations
- Connecting the systems you already run so data moves without manual re-entry.
How I work
- 1
Discovery
You get a written scope and a clear picture of the problem — before any code is written.
- 2
Prototype
You get something working you can actually use, not a slide deck.
- 3
Deliver
You get a deployed system with a proper handover — documentation, access, and a plan for what's next.
What I work with
- Languages & runtimes
- Python, TypeScript/JavaScript, Node.js — day-to-day tools for backend services, data pipelines and applied ML.
- ML & data
- PyTorch, scikit-learn, LLM-based extraction and retrieval — from classical ML to document AI.
- Platforms & infrastructure
- Cloud deployment, containers, CI/CD, relational and vector datastores — shipping and operating what gets built, not just prototyping it.
- Hardware-adjacent
- Embedded and electronics diagnostics, component-level repair, audio signal chains — the same rigor applied to physical systems as to software.