Technical Founder and Founding Engineer
Louis Forster
I build AI products from zero to real users. I can ship deep ML systems, close the product loop with users, and run venture experiments that quickly de-risk what to build next.
Current Focus
Looking for a founding engineer role (or technical co-founder fit) in an AI-first startup. Open to aligned pre-seed conversations.
Beta users
0+
On Treyspace
AI interactions
0K+
In production
Ventures
0
Built, shipped, validated
Selected Work
Projects below are ordered to show execution signal first: shipped product, deep technical research, and venture strategy.
AI workspace shipped to 26 beta users with 5,000+ interactions and 13M tokens processed. Built the full stack: GraphRAG engine, collaborative backend, tool-execution layer.
- Spatial whiteboard events map into typed graph entities with source provenance.
- Embedding retrieval returns context windows anchored to board coordinates.
- Tool execution is schema-validated, then merged through a state-safe session gateway.
Founder lesson
Pivoted B2B to B2C when enterprise adoption stalled. The product was good, but the channel was wrong. GTM has to be engineered alongside the system.
Bristol Vision Institute - Research
Reached state-of-the-art on the BVI-RLV benchmark (29.22 dB PSNR) by re-engineering a CNN-RNN video restoration pipeline on HPC infrastructure.
Technical Detail
Outcomes
- Higher temporal consistency
- Faster experiment cycles
- More stable tensor pipelines
Stack
Learnings
- Bidirectional warping stage for forward/backward feature alignment.
- ConvGRU temporal passes to stabilize frame-to-frame detail.
- Tensor-shape mismatch instrumentation across concatenation and recurrent blocks.
BECCS Innov - Venture Strategy
Ran the commercial feasibility for a carbon-credit supply chain venture. Killed assumptions before they killed the business: policy risk, buyer demand, unit economics.
Interview rounds
4
Prioritised bets
3 core
Market lens
TAM/SAM/SOM
Customer and operator interviews surface non-obvious constraints.
Double Diamond Process
Soft Skills & Strategy
Outcomes
- Validated go/no-go assumptions
- Mapped market entry wedges
- De-risked pilot sequencing
Stack
Learnings
- Iterative user feedback loops changed roadmap order and prevented overbuilding low-value features.
- Ruthless prioritisation kept the team aligned on only the assumptions that could kill the venture fastest.
- Clear team dynamics rituals reduced conflict and made decisions faster under uncertainty.
- Non-technical stakeholder feedback was translated into concrete engineering tasks with measurable acceptance criteria.
Open-Jet - Building in Public
Built in public with open roadmaps, weekly shipping, and community feedback loops to tighten product direction and distribution.
Release cadence
Weekly
Feedback source
Public threads
Iteration model
Ship -> learn
Push visible updates frequently so users react to concrete progress.
Building in Public
Outcomes
- Weekly shipping cadence
- Public roadmap discipline
- Community-informed feature prioritization
Stack
Learnings
- Shipping in public improves decision quality because assumptions are challenged early.
- Roadmap visibility creates accountability and forces clearer scoping discipline.
- Fast feedback from builders and users highlights edge cases before formal launch.
What I'm looking for
I am looking for a founding engineer or technical co-founder role where the core risk is product and execution, not just implementation.
I can own product and engineering end-to-end. My known gap is enterprise distribution at scale, so I work best with strong go-to-market counterparts.
I bring
- AI systems from research to production reliability
- Founder-speed full-stack shipping
- Direct user validation and rapid iteration
- MEng Computer Science, First Class (Bristol, 2026)
I'm open to
- Founding Engineer with end-to-end ownership
- Technical Co-Founder roles at pre-seed/seed
- Aligned investment to build the next venture