Meet Philip Kiely
Philip Kiely is a San Francisco Bay Area developer-relations leader, author, technical writer, and speaker focused on production AI systems. His work centers on inference engineering, model serving, GPU infrastructure, and helping developers understand how to deploy, optimize, and scale generative AI in production.
Summary of Philip Kiely and his portfolio site
Philip Kiely is a San Francisco Bay Area developer-relations leader, author, technical writer, and speaker focused on production AI systems—especially inference engineering: deploying, serving, and optimizing generative AI models at scale. His site presents him as an early Baseten employee (“employee number ten”) working in a broad role around developer relations, technical education, and AI infrastructure.
His current work centers on the practical engineering problems behind AI products: GPU infrastructure, model serving, quantization, latency/throughput/cost trade-offs, and frameworks such as TensorRT-LLM, SGLang, and vLLM. He describes inference as a young, valuable field where engineers work “from CUDA to Kubernetes” to make generative models faster, cheaper, and more reliable in production.
Main areas of work
-
AI infrastructure and inference engineering
Philip’s recent work is strongly focused on production model serving: quantization, speculative decoding, KV-cache optimization, multi-region GPU deployments, request prioritization, token accounting, and the “golden triangle” of latency, throughput, and quality. -
Developer relations and education
At Baseten, he publishes educational material, gives talks and workshops, and explains complex AI infrastructure topics for developers building production systems. -
Technical writing and books
Philip has a long-running technical writing background, with books, guest articles, documentation, blog posts, and conference materials. -
Career and communication advice
Earlier work includes technical writing for developers, authentic networking, and career-building advice for students and early-career technologists.
Books and notable projects
His Books page highlights three major publications:
- Inference Engineering — his latest book, published with Baseten Books, about becoming an expert in production AI inference. It draws from his Baseten writing, talks, interviews with engineers, and conversations with AI builders.
- Life-Changing Email — a guide to career-changing outreach and networking, aimed especially at students and early-career professionals.
- Writing for Software Developers — his first book, focused on helping software developers write technical content, build a portfolio, and improve communication.
Blog and writing
His Blog is a personal blog, while he notes that most of his current writing appears on Baseten’s blog. Recent personal posts include topics like building an AI audiobook, writing and designing Inference Engineering, using AI tools in DevRel, proofreading with AI, using Cursor to save time on a book project, and generating PDFs with Tailwind CSS.
He also maintains a Guest Posts page listing earlier technical writing across publications such as Smashing Magazine, CSS-Tricks, Twilio, WonderProxy, FloydHub, mkdev, Markforged, YotaScale, ContentLab, and Grinnell College. The older portfolio shows a broad technical-writing foundation in Django, AWS, localization, web scraping, cloud cost optimization, Git, and developer tooling.
Speaking and media
Philip has an extensive Speaking page with conference talks and workshops at AI Engineer World’s Fair, NVIDIA GTC, AWS re:Invent, PyTorch Conference, The AI Conference, ODSC, and other events. Topics include:
- high-performance inference for frontier AI models
- TensorRT-LLM optimization
- low-precision inference and quantization
- voice model inference with low latency
- SGLang and LLM serving
- embedding-based search and recommendation systems
- the economics of multi-model AI
His Media page lists many podcast and interview appearances, including Software Engineering Radio, Hanselminutes, Latent Space, TWIML AI, AI Engineering Podcast, Weaviate Podcast, CodeNewbie, and others. These appearances cover both his current AI infrastructure work and his earlier writing/creator journey.
Personal style and contact
On his Contact page, Philip presents a “standing invitation” to meet new people and talk about technology, business, sports, or interesting projects. He explicitly encourages students with evidence of unusual achievement, ability, or ambition to reach out for advice. He is clear that he is not looking for a new job, does not do freelance work, but may consider consulting, advising, investing, speaking, or podcast appearances.
Outside work, he describes himself as a lifelong martial artist, avid reader, and enthusiastic Bay Area sports fan.
Social profiles
Philip links to these public profiles from his site:
Overall, the site positions Philip Kiely as a technical communicator and developer-relations leader whose work has evolved from broad software-development writing into deep specialization in AI inference, production model serving, and developer education for AI infrastructure.
- None found