Graviton Research Capital is one of the largest quantitative trading firms emerging out of APAC, with a robust footprint across global exchanges and a rapidly growing presence across Europe and the USA. We’re long-term focused, and deeply invested in building high-performance teams that operate at scale in some of the most competitive markets globally. You will be part of the team that builds and manages AI platforms.
What you’ll do:
- Deploy and operate open-weight LLMs on our internal servers.
- Build and run an internal LLM gateway/router between in-house and external models.
- Size and manage internal infrastructure capacity across teams: scheduling, monitoring and utilization.
- Tune serving for throughput and latency: KV cache management, batching, quantization and context-length trade-offs.
- Evaluate new models and inference techniques as they appear, and advise on what to adopt.
- Support fine-tuning of models on internal data and help evaluate the results.
- Help teams use the platform with coding assistants, internal tools, retrieval over documentation.
What We’re Looking For:
- 7+ years of engineering experience, preferably in Python.
- Hands-on experience deploying and serving LLMs on on-prem servers.
- Experience with or strong working knowledge of LLM gateways/proxies (LiteLLM, OpenRouter or in-house).