Bio
Accomplished engineering leader with 15+ years of experience in AI, cloud-native platforms, and infrastructure. Proven track record of building and scaling high-performing teams and delivering significant performance improvements in enterprise AI products. Combines deep technical expertise in AI/ML with strategic vision to drive product innovation and business impact.
Session (1)
Additional Speaker Details
Using publicly available information we constructed an analysis to help you get a feel for this speaker before deciding to attend their session.
Ashish Kamra is Senior Manager of AI Performance at Red Hat, where he leads inference performance work on the open-source vLLM serving stack and is first/lead author on Red Hat's MLPerf Inference v6.0 results across NVIDIA and AMD hardware. His session offers production lessons on maximizing throughput and efficiency from open-source LLM serving at enterprise scale.
Speaking style
GitHub
@ashishkamra- ashishkamra/vllm-memory-estimator★ 1GPU memory estimator for vLLM: parses serve commands to give per-GPU memory breakdowns and budgeting for context lengths and concurrency
- ashishkamra/AI_usage_analysis★ 0AI usage survey analysis tool with semantic categorization using the Phi-4 model and interactive visualizations
Recent writing (4)
- Red Hat and NVIDIA: Setting standards for high-performance AI inference · blog · Apr 2026
- Red Hat AI tops MLPerf Inference v6.0 with vLLM on Qwen3-VL, Whisper, and GPT-OSS-120B · blog · Apr 2026
- 5 steps to triage vLLM performance · blog · Mar 2026
- Efficient and reproducible LLM inference with Red Hat: MLPerf Inference v5.1 results · blog · Oct 2025