Fuse Energy ir Apvienotās Karalistes enerģētikas uzņēmums, kas koncentrējas uz to, lai elektroenerģijas piegāde būtu vienkāršāka un pārskatāmāka. Tā lietotne palīdz klientiem mainīt elektroenerģijas piegādātāju un piekļūt reāllaika norēķinu informācijai, savukārt uzņēmums piedāvā konkurētspējīgus elektroenerģijas tarifus un attīsta saules un vēja enerģijas projektus. Fuse Energy arī novirza peļņu starptautiskām atjaunīgās enerģijas iniciatīvām, apvienojot praktiskus enerģētikas pakalpojumus ar plašāku uzmanību tīras enerģijas attīstībai.
Applied AI Researcher - Fuse Energy (London Hybrid)
Develop and evaluate dependable AI systems for Fuse Energy’s solar, battery, grid, trading, and home-energy operations.
Apraksts
- Design experiments that identify the most effective model and approach for each workload while balancing accuracy, latency, and cost
- Own the evaluation framework, from high-quality datasets and golden answers to human and LLM judgments and evaluator-confidence measurement
- Investigate post-training techniques such as supervised fine-tuning and reinforcement learning
- Build agent harnesses for dependable long-running tasks, incorporating tool use, deterministic checks, state tracking, retries, resource allocation, and human approvals
- Define evaluation and escalation thresholds based on the consequences of failure
- Collaborate across Fuse to turn research findings into production systems and introduce frontier applied-AI methods throughout the company
Prasības
- Demonstrated experience applying modern AI models to real-world problems and improving measurable outcomes through experimentation
- Strong knowledge of evaluation design, dataset quality, and failure analysis
- Practical experience with model selection and routing, fine-tuning, reinforcement learning, agent systems, or computer-use systems
- Strong software engineering capability, including Python and the development of dependable experimental and production tooling
- Sound judgment when balancing cost, latency, reliability, and risk, including decisions about deterministic checks and human review
- Able to move fluidly from research to implementation by forming hypotheses, running rigorous tests, examining failures, and shipping effective solutions
- Experience in a frontier AI lab or a similarly experimental team is particularly valuable
Priekšrocības
- Competitive salary with equity eligibility
- Bonus scheme paid twice a year
- Fully funded technology tailored to your needs
- Private health insurance
- Breakfast and dinner allowance for employees working from the office
- Benefits may differ by location