The central challenge is reliability. A strong demo is not enough: agents must function across different devices, operating-system versions, aspect ratios, network conditions and inconsistent loading behaviour. They need to recognise when something has gone wrong, recover intelligently and continue operating without supervision.

Your systems will test real products, identify real bugs and influence how an entire industry approaches quality assurance.

What You’ll Be Doing

• Build end-to-end agents that connect visual perception and game state to concrete actions such as taps, swipes and clicks.
• Design robust observe–decide–act–recover loops rather than brittle, pre-recorded scripts.
• Develop cross-device automation that works across mobile and desktop environments.
• Enable complex user journeys to run across different devices, operating systems and aspect ratios without depending on fixed coordinates or hard-coded timings.
• Automate critical flows including tutorials, purchases, onboarding and core gameplay loops.
• Work with low-level video streaming, device infrastructure, automation platforms and infrastructure as code.
• Instrument, debug and improve agents operating under slow networks, variable load times, unstable SDKs and other real-world conditions.
• Establish internal evaluations and production standards for agent performance and reliability.
• Work directly with QA and engineering leaders at game studios to understand their workflows.
• Translate ambiguous customer problems into dependable automation and influence the product roadmap.

What We’re Looking For

• Experience building vision-driven or multimodal agents using video—not solely text or static images.
• A strong computer-vision foundation and familiarity with Vision-Language Models or related multimodal architectures.
• Experience delivering meaningful automation in messy, stateful environments such as mobile applications, browser automation, robotics, autonomous systems, device farms or complex user interfaces.
• An understanding of closed-loop agent systems that observe, reason, act, validate outcomes and recover from failure.
• Strong systems-engineering instincts and an obsessive approach to reliability, instrumentation and debugging.
• Experience owning automation that other teams rely on in production.
• The ability to work comfortably across AI models, control logic, infrastructure and product engineering.
• Confidence working directly with customers and converting loosely defined operational pain into practical technical solutions.
• Experience in a zero-to-one AI product environment, or a strong appetite for the ownership and ambiguity that comes with one.

An interest in games or other complex interactive systems is helpful, although deep gaming knowledge is not required.

Why Join?

• Build autonomous agents that take real actions in live environments.
• Solve difficult problems across applied AI, automation, infrastructure and systems engineering.
• Own systems from initial architecture through to production reliability.
• Work directly with founders, customers and engineering leaders at major studios.
• Influence both the technical foundations and product direction of an early-stage company.
• Join a high-trust environment where strong ideas can ship quickly.
• Receive meaningful equity and share in the value you help create.
• Flexible working hours and a culture focused on impact rather than process.

Please apply directly here or email your CV over to adam.lockett@sr2rec.co.uk