Skip to content
View topppg's full-sized avatar
📕
📕

Block or report topppg

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
topppg/README.md

Giorgio Sarno

AI Software Engineer at Replika

My work centers on large language model inference, from serving models efficiently to planning GPU capacity. I also build infrastructure for AI agents, covering tool use, background tasks, and coordination across multiple agents, and I contribute to a memory layer that supports them.

Before this, I worked in data platforms and production machine learning. I designed and deployed custom embedding models for matching industrial parts, and I led a team that brought time series models into production for fleet monitoring. Earlier still, my PhD research in Marseille explored numerical approaches to quantum gravity. Along the way I wrote open source C libraries, including sl2cfoam, with CUDA kernels for GPU parallelization.

Experience

  • Replika AI · AI Software Engineer (Oct 2025–present): LLM inference with vLLM and SGLang, GPU sizing, API routing, agent infrastructure, and memory systems.
  • Sparrow Parts · Head of Data & AI (Feb–Oct 2025): custom embedding models for industrial part matching and direction for the data platform.
  • Mate · Founding AI Engineer (2024–2025, part-time): an event driven recommendation system, cold start matching, and behavioral signals.
  • Stratio · AI Team Lead (2022–Feb 2025): production time series models for fleet monitoring and automotive diagnostics.
  • Knots Analytics · Co-founder (2020–2023): route optimization for the marine industry and product and technical direction.
  • Centre de Physique Théorique de Marseille · Research Scientist · PhD (2017–2020): numerical quantum gravity research and open source C libraries with CUDA kernels.

Technical skills

  • LLM serving: vLLM, SGLang, KV caching, LLM evaluation
  • Agents and retrieval: agent orchestration, MCP, memory systems, RAG
  • Machine learning: PyTorch, embeddings, recommendation systems, time series models
  • Engineering: Python, FastAPI, Kubernetes, Docker, C, CUDA

Links

Website · LinkedIn · X · GitHub

Pinned Loading

  1. spiritboard spiritboard Public

    Your agents. One board. Open it and immediately know which agents need your attention, what they're doing, and where to continue.

    Python 1

  2. gauge-attention gauge-attention Public

    A PyTorch research prototype for learned position-dependent rotary transport in attention

    Python

  3. sparse-decode-bench sparse-decode-bench Public

    Training-free page-sparse decode in llama.cpp on consumer GPUs: pre-registered speed and RULER results

    Python