About Grantx
We're a start-up, founded in 2024, building the intelligence layer for the $1.7 trillion US grants market. Our ML stack achieves accuracy scores that are 5x the market leader, developed for a fraction of what competitors spent on their systems.
We're not wrapping foundation models and calling it AI. We're doing real research, shipping real infrastructure, and proving that a small team with the right architecture can outperform and capture the market.
9 people. Post-GA and scaling into enterprise. If you've been waiting for ML work that actually matters, this is it.
The Role
You'll work directly on our core engine, a proprietary hybrid system that combines neural and symbolic reasoning in ways that are genuinely non-obvious. The architecture was designed from first principles and is now the foundation of everything we ship.
You'll be advancing a system that's already setting new benchmarks in information retrieval, pushing it toward a fully neurosymbolic stack while keeping inference fast and production-grade.
You'll Be Working On:
* Advancing retrieval and reasoning architecture, neural, symbolic, and the space between
* Domain adaptation and custom training pipelines across a corpus of 650K+ specialised documents
* Serving infrastructure: low-latency inference at scale on GCP
* Verification and evidence systems: making ML outputs explainable and defensible
* Entity resolution, graph intelligence, and signal extraction across millions of records
* Evaluation frameworks that benchmark against the best in the market, and win
What We're Looking For:
* Strong foundations in information retrieval, NLP, or applied ML (you should understand the math, not just the APIs)
* Experience with embedding models, dense retrieval, or hybrid search systems
* Comfortable going from research paper to production code without a separate team doing the translation
* Python. GCP familiarity helpful (Spanner, GCS, or equivalent cloud-native tooling)
* Genuine curiosity about neurosymbolic approaches, if you've read Garcez or Marcus and got excited, we should talk
* Pragmatic. We ship weekly. Perfectionism without velocity doesn't survive here
The Non-Negotiables:
* UK-based
* Minimum 4 hours daily overlap with US Eastern Time (1pm–5pm GMT as a baseline)
* Our product, data, and leadership teams run on EST. Standups, design reviews, sprint coordination, you need to be in the room
What You Get:
* Competitive salary + meaningful equity in a company with a defensible technical moat.
* A small, intense, brilliant team where your work ships to real users the same week
* The chance to publish or patent work that's actually novel.
The Vibe:
This is skunkworks energy. Small team, big ambition, zero bureaucracy. We built something that embarrasses systems 100x our budget. Now we need someone who wants to take it further and isn't afraid of the pace.
If you want a well-defined role with clear role boundaries and a gentle ramp, this isn't it. If you want to do the best ML work of your career on a system that's genuinely pushing the field, get in touch.
Job Type: Full-time
Pay: £65,000.00 per year
Benefits:
* Casual dress
* Employee stock ownership plan
* Gym membership
* Private medical insurance
* Unlimited paid holidays
* Work from home
Application question(s):
* Why you?
Education:
* Bachelor's (required)
Language:
* English (required)
Work authorisation:
* United Kingdom (required)
Work Location: Hybrid remote in London WC1V 6LJ