Quham Adefila

Technical founder and operator

You sell it.
I’ll build it.

Agentic trading systems used by PMs to manage billions in exposure. From stalled to successful deployment worth half a startup’s revenue. My own company, built from zero. Now I’m looking for an exceptional commercial co-founder to build something much bigger with.

Get in touchView CV

Where I’ve built

01 Highlights

Value and execution.

Deep research. High-stakes production systems. LLM and agentic workflows. I’ve built all three.

run·track_record

4/4 delivered · awaiting input

✓ systems

Rokos

Agentic trading systems and core AI leadership.

BillionsIn exposure

Built an agentic trading system used by the firm’s founder and equity PMs to adjust billions in exposure. Led the core AI team and overhauled the internal data structure so Claude could access the firm’s knowledge.

From PMs to HR, people across the firm now build and deploy their own applications.

Production systemsAI adoptionData infrastructure
✓ execution

Scope

Shipped a deployment stalled for nine months.

~50%Company revenue

As a founding engineer, led the successful initial deployment of a project that had been stalled for nine months.

The project accounted for roughly half the startup’s revenue. I quickly got up to speed in the inspection industry to carry it through to production.

Product deliveryTechnical ownershipCommercial execution
✓ research

DeepMind

Typhoon forecasting with video models.

30%Benchmark improvement

Owned ML pipelines for one of the hardest prediction problems: typhoon forecasting with video models. The approach improved on existing benchmarks by 30% at a 14-day forecast horizon.

Machine learningVideo modelsWeather forecasting
✓ distribution

Inspire Maths Tuition

Team of 10, six-figure revenue and millions of TikTok views.

10Team members

Founded and scaled Inspire Maths Tuition alongside university, growing the business to a team of 10 and six-figure revenue, with millions of views on TikTok.

Customer acquisitionLeadershipBusiness operations
next

A commercial cofounder to build pipeline and lead fundraising.

Let’s talk

02 Project gallery

The work, in context.

Systems, experiments and businesses.
A closer look at the work behind the results.

Planner delegates financial research to parallel agents across data sources, then synthesises a cited thesis.

Multi-agent research engine

Financial intelligence · Designed & built

Pre-earnings brief workflow comparing internal investment research with market consensus, verifying facts and preparing a portfolio-manager report.

Pre-earnings brief

Financial intelligence · PM decision support

Scribe AI meeting analyser interface showing a live transcript, linked research and analyst chat.

Scribe

AI meeting analysis · Research graph

Africoin concept dashboard for an inflation-adjusted, reserve-backed stablecoin proof of concept.

Africoin

Reserve-backed stablecoin · Solo POC

Civil-impact planning interface showing synthetic demonstration data, marked not operational.

Civil-impact picture

Geospatial planning · Synthetic demo

Typhoon forecast track and benchmark visualization from a video-model research project.

Typhoon forecasting

Google DeepMind · Research

Suitify generative AI product for creating professional headshots from a photo.

Suitify

Consumer product · Generative AI

Inspire Maths Tuition tutor directory and education marketplace website.

Inspire Maths Tuition

Founder · Product & distribution

03 Engineering mindset

Engineering is now orchestrating the agent loop.

Designed to iterate in one specific problem space.

AGENT LOOP

01 / Context

The right information.

The model can see the data and knowledge that actually matter.

Model providers will eventually build this whole loop in. That’s a hard problem and it won’t happen soon.

Until then, whoever designs these loops best for a domain wins it.

04 The team I’ll Build

Curious minds.
Commercial instincts.

The people I want around me care about the outcome.
They take responsibility for getting there.

/ 01

Own the outcome.

Care about what the customer achieves. Follow the work through to something useful.

/ 02

Question the default.

Understand why a process exists. Ask whether it should. Be willing to change your mind.

/ 03

Think commercially.

Connect technical choices to demand, distribution and the economics of the business.

Code is cheap.
Good judgement isn’t.

05 The next chapter

Exploring commercial cofounder partnerships

Something big.
Built together.

I’m looking for a commercial cofounder.

Someone who can build customer pipeline, turn relationships into revenue and lead fundraising.

I bring technical depth and execution across AI, systems and product. Together, we own the strategy, the business and the outcome.

The right two people together get far more done than either alone.

Startups move fast, and AI is accelerating the pace. I’m looking for someone who:

  • Adapts and executes quickly. Keeps us ahead as the market shifts.
  • Leads through highs and lows. Keeps the team steady when things get hard.
  • Reads people well. Understands customers, future hires and investors.
  • Shares the ambition and chemistry. Building something massive is a marathon; we should enjoy running it together.
Get in touch