Free LLM Inference Engineering Study Group: From KV Cache to vLLM
V-on-run4_here · reddit · 2026-11-16
A Reddit user is organizing a free study group on LLM serving and inference engineering, aimed at beginner-to-intermediate levels and focused on first-principles understanding of LLM inference and GPU optimization.
Topics covered:
- Transformer architecture fundamentals
- KV Cache and memory bottlenecks
- PagedAttention and virtual-memory principles in LLM hosting
- Continuous batching for throughput and iteration-level latency
- vLLM architecture overview
The organizer plans to set up a WhatsApp group and coordinate cross-timezone meetings once at least 10 participants join, stressing it is free with no paid follow-up course and not a promotion; the poster describes themselves as a junior engineer, not a mentor.
More from Infra
- OpenAI made ~$13B last year but committed ~$1.15T to compute — Qweskj · 2026-11-16
- Pirateface launches P2P network for sharing open models without centralized hosting — alohahaho · 2026-11-16
- China, not the US, is blocking H200 sales; analyst calls chip policy decade's biggest self-own — firstadopter · 2026-11-16
- Agents need throwaway DB copies, not mocks — copy-on-write cloning may become default infra — petrusenko_max · 2026-11-16
- Self-hosted daily game pipeline runs Qwen3.6-35B and Qwen-Image locally on DGX Spark, zero API fees — jbsan · 2026-11-16
- audio.cpp ships real-time local avatar voice chat: ~850ms latency on RTX 5090 — Acceptable-Cycle4645 · 2026-11-16