// learn together
Every meetup is part of a structured curriculum. Sessions are technical, progressive, and engineering focused—not general networking events.
Join us on Meetup →// every session
// topics we go deep on
We don't do introductory overviews. Every IMLG session targets engineers who want to understand how things work at the systems level.
Streaming multiprocessors, Tensor Cores, HBM, NVLink and NVSwitch topologies
CUDA programming model, Triton, and the NVIDIA ecosystem — cuBLAS, NCCL, NIXL, cuDNN
Intra-node and inter-node communication: PCIe, GPUDirect, RDMA, DPU
Data, model, pipeline, and tensor parallelism; FSDP; Megatron-LM; torch.compile
Experiment tracking, feature stores, vector and graph databases, serving pipelines
KV cache management, speculative decoding, PEFT, RLHF strategies, inference optimization
DCGM, Nsight, OpenTelemetry, Prometheus, ClickHouse, Grafana
OpenLLMetry, Langfuse, Phoenix, TTFT and ITL SLOs
Linux administration and platform engineering for GPU infrastructure
Edge (Jetson, NPU), on-prem GPU servers, and cloud GPU instances
Join our community to receive announcements, session materials, and recordings from every IMLG meetup.
Join IMLG on Meetup →