Software Engineering Manager
About the job
- Fluency with artificial intelligence and large language models, including LangChain, vector stores, and prompt engineering.
- Full-stack capability across frontend, backend, and infrastructure.
- Experience shipping product in a startup environment with little or no safety net.
- Relentless drive and ability to execute and adapt quickly.
- Product orientation focused on outcomes.
- Experimental instinct to build, test, and learn iteratively.
- Own engineering execution end-to-end.
- Take ideas from inception to production.
- Help define the company's technical direction.
- Deploy daily and learn rapidly.
- Talk to users and shape the roadmap, including working with founders on calls that determine what to build.
- Train AI agents by designing large-language-model-powered testers that explore, learn, and adapt in real time.
- Own the technology stack across Python, TypeScript, and cloud infrastructure, from commit to production.
- Run experiments on models, embeddings, and retrieval pipelines, and turn prototypes into production.
- Ship AI features in production and build products from zero to one.
- Experience shipping AI features in production.
- Experience building something from zero to one.
- Experience thriving in a chaotic environment with high ownership.
About the company
Duku AI provides a SaaS platform for automated exploratory testing that checks if a web product truly works for users, not just whether code runs or unit tests pass. Users point Duku at a URL, and the system performs real-flow explorations to surface silent breaks such as broken dropdowns, failed state transitions, and forms that submit nothing. Unlike tests that focus on code or API responses, Duku demonstrates end-to-end product behavior and catches issues that affect user experience, regardless of whether the code comes from engineers, Copilot, or other LLMs. Engineering teams use it to detect revenue- and trust-impacting product failures before customers encounter them. The goal is to provide continuous, automated product exploration and failure detection to ensure websites behave correctly for real users.
Company Size
11-50
Company Stage
Seed
Total Funding
$160K
Headquarters
London, United Kingdom
Founded
2025
Get referred to Duku AI
See people who can refer or advise you
Simplify's Take
What believers are saying
- Duku reported 146K autonomous runs and 3K product surfaces by July 2026.
- The company shipped seven major releases in June and July 2026.
- A GitHub Action integration fits engineering workflows and shortens adoption friction.
What critics are saying
- Duku raised only GBP 120,000 from Antler in August 2025, limiting runway.
- Built In showed 14 employees in May 2026, making enterprise sales and support thin.
- QA Wolf and similar vendors sell deterministic Playwright output, undercutting Duku's agentic pitch.
What makes Duku AI unique
- Duku turns URLs into autonomous user-flow maps, not brittle scripted Playwright suites.
- Its pitch targets silent UI failures, like broken dropdowns and dead form submissions.
- July 2026 docs stress replaying critical journeys on every deploy, not after release.
Help us improve and share your feedback! Did you find this helpful?