// working_with_me

How we’d work together.

I take on a small number of fractional and advisory engagements at a time, alongside my work at Mantis — remote, across UK / EU / US timezones. Most start with a fixed-fee audit or roadmap to build trust and a plan, then convert to an ongoing fractional retainer. Pricing reflects senior, domain-specialist AI leadership — a decade of shipping, deep LLM work, and a fintech edge — not generalist consulting.

$ book a 30-min intro call_ email me

01 AI Audit & Roadmap Front door from £8K / 2–5 wks

A fast, senior read on where AI can actually move the needle for you — and where it can’t — through to the plan to get there. Scoped to what you need: a capability audit (opportunity map, data / readiness / risk, first build-vs-buy calls, an exec-ready deck), or the full 12-month roadmap (architecture, model & vendor strategy, an evaluation framework, and a team & hiring plan, with governance and cost guardrails). Either way you leave with something you own and can take to the board. Credited against a retainer if you go on to one.

Prioritised opportunity map
Data / readiness / risk assessment
Build-vs-buy calls
12-month roadmap
Architecture & evaluation strategy
Hiring plan & guardrails

// where this shows up

The Heritage Fund. For the UK’s largest heritage funder, this looked like leading a workshop on AI and business transformation for their leadership, then scoping AI tagging and classification across their grants — assessing where AI genuinely helps, aligning it to the mission, and setting the plan before anyone builds.

“Calm, collected, and always on top of the ML/AI industry — steering the ship with precision without falling for hype. He aligns technical decisions with clients’ business objectives, and as a leader provides valuable guidance and takes feedback into account.”

Raphael Mitsch AI Team Lead, Climatiq

02 Fractional Head of AI core offer £12–18K / month

Ongoing, embedded AI leadership — one to two days a week, at a fraction of a full-time Chief-AI-Officer hire. I own the AI strategy and its delivery end to end: set direction, lead and hire the team, make the vendor and architecture calls, report to your board as the accountable AI voice, and stay hands-on where it counts — the hard evals, the key builds, the decisions that are expensive to get wrong. Three-month minimum; cash, or cash + a small equity blend for venture-stage teams.

Own strategy + delivery
Lead & hire the team
Vendor / architecture decisions
Board reporting
Hands-on where it counts

// where this shows up

Datamaran, Addgene & Oxford Medical Simulation. Embedded, hands-on delivery. Built and deployed a retrieval-augmented (RAG) system — with its own evaluation harness — into Datamaran’s ESG-intelligence platform. Currently own the AI recommendation-engine build and its expert evaluation for Addgene’s research catalogue. Ran an embedded ML team inside Oxford Medical Simulation’s own stack for their VR clinical training. Owning the strategy, the build, and the evals — not just advising.

“Mantis have been a key component of our growth within the LLM domain. Their expertise, pragmatism and collaborative approach has made them a trusted partner to accelerate our ML advancements.”

Michael Wallace CEO, Oxford Medical Simulation

03 AI Advisor / Board Light-touch from £3K / month

Judgment on tap. A monthly strategy session plus async access — sanity-checks on hires, vendors and roadmap, and an experienced AI operator in your corner for the calls that are easy to get wrong. Board seats include a small equity grant. One-off expert or diligence calls also available.

Monthly strategy session
Async access
Hire / vendor / roadmap sanity-checks
Board seat option

// where this shows up

GSMA & FIND. High-leverage, bounded work. For GSMA, designed the Q&A dataset and open-sourced the dataset-creation tooling behind their Open Telco AI models, with a contribution to the accompanying NeurIPS paper. For FIND, built a multi-LLM extraction pipeline — benchmarking GPT-4, Claude and Gemini against a human-labelled ground truth — to de-risk an LLM approach before they committed to it. The judgment to scope the right thing, and prove it works.

“We trust the quality of Mantis’s work and the diligence they take in getting to solutions. They work collaboratively, support our internal data scientists, and feel like part of our team.”

Alasdair Fraser Head of Data & Digital, The Wellcome Trust

[→] // let’s talk

Start with a call

If one of these fits — or you’re a founder or executive trying to make real AI decisions, not buy demos — the fastest start is a 30-minute intro. No pitch theatre; we’ll see if it’s a real fit.

> ssh nick@nsorros.com

  email   nick.sorros@proton.me
  call    meet.nsorros.com
  in      linkedin.com/in/nsorros

  status  available q4 2026

$ compose --to nick.sorros@proton.me_