Computer Vision Engineer

**Confidential — Independent Cricket-Trading Firm** Computer Vision Engineer; £100K–£110K + quarterly bonus; London (Hybrid)

**About Company** An independent, profitable quantitative sports-trading company — algorithmic trading applied to sports betting markets. This side of the business is focused exclusively on cricket, where the company has spent the last decade building into one of the established leaders in cricket trading. The company has no external funding and is entirely self-financed.

The company is reinvesting in computer vision and data to expand what feeds its trading models. The deliberately understated JD reflects a small, tight-knit, competitive industry where the company prefers to keep its hiring quiet. Culture is opinionated and fast-moving: people who thrive here are confident, proactive "doers" who speak up, propose fixes, and are comfortable wearing multiple hats, often with startup backgrounds.

**About the Role** This is the company's first dedicated computer vision hire. A contractor built the initial pipeline extracting data from cricket broadcast footage; this person takes full ownership of it, becoming the sole CV specialist at the company. They sit functionally with the Data team, reporting to the Lead Data Engineer, with the Head of Data as the internal stakeholder.

**Responsibilities** - Take ownership of the existing computer vision pipeline (built by a contractor) extracting data from live and historical cricket broadcast/replay footage. - Develop and improve models across object detection, tracking, pose estimation, action recognition, and/or temporal event detection across variable camera angles and broadcast-quality footage. - Prevent double-counting or false capture of events in replayed segments. - Build validation frameworks checking automated extraction against manually-collected ground truth from the Cricket team. - Define metrics, tests, and tooling to keep extracted data accurate and reliable at scale. - Pick up existing in-flight workstreams and priorities on the current pipeline.

**Required Skills** A. Proven experience building and deploying computer vision models in production, end to end. B. Demonstrated depth in at least 2–3 of: object detection, tracking, pose estimation, action recognition, temporal event detection. C. Strong Python for production ML systems. D. Comfortable working autonomously and owning technical decisions without a CV team around them.

**Bonus Skills** - Prior sports or broadcast-video experience (nice-to-have, not required — current contractor has none). - Background at a sports-tracking company (e.g., Hawkeye) or similar; general sports-tech experience valued over gambling-industry experience.

**Logistical Info** - Location: London (Hammersmith); firm 3 days/week in-office — Monday, Tuesday, Wednesday. - Compensation: £100K–£110K base + quarterly bonus (larger year-end payout); cash only, no equity/profit share. - Number of openings: 1 (client would consider a second hire within ~6 months depending on pipeline progress). - **Other:** Visa sponsorship considered for the right candidate.

**Ideal Background** - Strongest signal: CV engineers from sports-tracking/sports-tech companies (e.g., Hawkeye) — sports-tech background valued over gambling-industry background. - Client identity (confidential — do not share with candidates until 1:1 conversation): Mustard Systems, mustardsystems.com. - Confidential search: don't name the client in outreach until candidates reach 1:1 conversation — client wants this hire kept quiet from competitors. - Headcount note: 1 opening now; possible second hire in ~6 months — don't oversell as a team build. - Skill-depth note: client explicitly wants a spike in 2–3 specialties, not broad shallow coverage.

**Green Flags** - Confident, opinionated "doer" — happy to work independently, speak up, push back. - Comfortable being the sole CV person, owning the roadmap without a team. - Startup background, used to wearing multiple hats.

**Red Flags** - Needs a structured team/manager environment — this is a solo CV seat. - No appetite for ambiguity — inherits in-flight priorities rather than a blank slate.

**Interview Process** 1. Intro call (~20 min) with the Head of Data. 2. Take-home technical test (~1 week; still being finalized by the client). 3. In-person interview with the CTO, Lead Data Engineer, and Head of Data.