TracePilot
Active buildAn AI-assisted incident-triage system using a Go API service and Python LangGraph worker to investigate failed software and data workflows and produce structured reports.
- Go
- Python
- LangGraph
- LLM workflows
Jacob Bickerstaff · Product-minded Software Engineer
Building reliable systems from ambiguous product problems.
I like understanding the problem before reaching for the implementation: finding the simplest useful path, then building backend, data, and AI-adjacent systems that hold up in real usage.
Best fit
Early-stage teams building data-heavy, backend-heavy, or AI-adjacent products.
How I work
I turn unclear problems into small, testable systems that can ship.
Strength
Balancing product judgement with backend correctness and delivery discipline.
Selected Projects
A small set of projects that reflect how I think.
These are product and engineering exercises with a common bias toward explicit system behaviour, deliberate trade-offs, and practical implementation.
Deterministic Document Resolution
Technical projectA deterministic document-analysis system for resolving acronyms, defined terms, and structural references in complex documents.
- Python
- FastAPI
- Pytest
- NLP
PocketBank
Active buildA Go backend project modelling banking operations through service/store boundaries, staged persistence, ledger-style records, and idempotency-aware writes.
- Go
- net/http
- PostgreSQL
- pgx
Clarity on Tax
Live productA public-facing UK tax calculator built with Next.js and TypeScript to make salary, pensions, dividend, National Insurance, and company extraction trade-offs easier to understand.
- Next.js
- TypeScript
- React
- Tailwind CSS
Positioning
I like working where the problem is still taking shape.
I’m interested in teams where engineering sits close to product and or the client: understanding the problem, finding the smallest useful version, and building systems that can grow without becoming a black box.
How I work
- Turn incomplete requirements into concrete delivery steps.
- Keep early versions small enough to ship and learn from.
- Build backend and data systems that are simple enough to understand, but structured enough to maintain.
Where I’m useful
- Turning prototypes into reliable services and workflows.
- Adding tests, deterministic checks, and observability around AI-adjacent systems.
- Bridging product ideas, backend implementation, and operational reality.
Engineering approach
Using AI tools without handing over judgement.
I use AI tools for scoped engineering work, but treat the output as something to review, test, and adapt — not something to trust blindly.
Where it helps
Exploration, implementation drafts, refactoring options, test improvements, and reviewing alternative approaches.
Where judgement stays
System design, correctness, maintainability, trade-offs, and deciding whether a change belongs in the codebase.
Contact
Interested in working together?
Interested in my work, projects, or a potential role? The best place to start is by email.
Email is the best way to reach me.
I’m based in Leeds, open to remote-first roles, Manchester hybrid working, and planned UK-wide team meet-ups.