Head of Research And Development

About SimplyfAIWe build Al to be relevant and reliable-then revolutionary. We love frontier research most when it lands as a working product. In a space rife with smoke machines and buzzword bingo, simplicity is the ultimate sophistication.Who we're looking forA builder-researcher who can find, test, and embed novel ideas into new and existing products. You'll shape how we design and use agentic systems, lean prototyping, and well-thought-out optimisation. You'll work closely with Engineering to ensure scalability, observability, and safety-without turning research into ceremony.In this role you willResearch.Scan, prioritise, and test frontier work (reasoning tools, routing/memory, small-model tuning) with tight experiments that prove value.Use pragmatic frameworks when they help (e.g., DSPy flows, MCP-style tool servers/function calling) — keep the stack simple.Nurture the research culture: run short reading groups, publish internal notes and public blog posts, host occasional community sessions; turn learnings into reusable kits and examples.Write crisp docs: what we tried, what moved the needle, what ships next.Product & RoadmapTie R&D to product goals: quality, latency, reliability, and cost.Partner with Engineering to take POCs to production-clean interfaces, tests, tracing, feature flags, rollbacks.Make sensible trade-offs (cache vs compute, model size vs accuracy, CPU/GPU) and keep systems observable and debuggable.Evaluation & OptimisationBuild eval harnesses that mirror real use: offline suites, honest A/Bs, and steady telemetry.Improve outcomes with practical optimisation loops (prompt/demо tuning, routing, budgeted inference) and sound statistics.Treat optimisation at both levels: system-level (data quality, caching, routing, cost/latency) and task-level (accuracy on specific workloads).Safety & StewardshipBake in privacy, data handling, guardrails, and red-teaming from day one-small rules, clear logs.Mentor a compact, senior R&D group; raise the bar on code, docs, and decisions.You might thrive in this role if you areIndependent and self-driven-motivated by solving real problems, not process theatre.Community-minded: you share what you learn (notes, talks, posts) and contribute back when practical.Proven history of picking the right tool for the job (from open-source models to agentic frameworks) and walking away from the wrong ones.Strong on fundamentals: probability, statistics, optimisation, linear algebra-shown through working systems and careful experiments.Credential-agnostic: a relevant degree is great; equivalent proof of deep technical understanding is even better.