Systems, products, and projects
The scale runs from a three-person school site to storage underneath large AI workloads. In every case, I start with the mechanism, build what is missing, and keep the result testable.
2017–2025 · Microsoft Azure Storage
I led Azure Blob product work and strategic engagement for demanding AI workloads, including the Azure-side storage relationship supporting OpenAI and Microsoft AI. I translated checkpointing, model distribution, namespace, throughput, capacity, and reliability requirements into product priorities and operating mechanisms.
OpenAI’s 2019 Dota 2 report names me under Microsoft Azure Support. Microsoft separately documents OpenAI’s exabyte-scale Blob account and 10-terabit-per-second data path. At SNIA SDC 2024, Jegan Devaraju and I explained how Scaled Accounts, Blobfuse, hierarchical namespace, and checkpoint movement supported large-scale model training.
My part was infrastructure and product leadership underneath the models, alongside a much larger team.
2024–present
AI can propose a change. Firmware still controls the greenhouse.
I am building an AI-assisted greenhouse control loop with my adult son, James.
The ESP32 state machine owns safety and actuation. AI can tune only bounded parameters, and telemetry plus a decision ledger tells us whether each change worked.
2024–present
A private developer platform that has to keep working.
I run a pod-per-repository fleet of coding agents on a live Kubernetes cluster.
GitOps defines the desired state. Review gates, model routing, observability, and durable worktrees give the agents room to work without giving up control.
2017–2025
The models still needed a storage system underneath them.
I led Azure Storage product work for demanding AI workloads, including the Azure-side OpenAI relationship.
I translated checkpointing, model distribution, namespace, throughput, and reliability requirements into product priorities and operating mechanisms. OpenAI later named me under Microsoft Azure Support.
2015
A small public artifact that anyone can inspect.
I contributed a 32-disk VM and storage-pool Azure Resource Manager template.
The original pull request added five files and 614 lines, and the README follow-up also merged. Azure revised the code later, while the successor metadata still credits jvallery.
2007–2016
I built the utility I needed, then turned it into a product.
MineMyMail extracted, deduplicated, enriched, and exported contacts from IMAP mailboxes.
It began as a Gmail address-extraction tool and grew into a paid product. It was small and imperfect, but people were still using it in 2016.
2006–2008
A working social-news and RSS experiment.
I launched and operated a Pligg-based site with accounts, submissions, voting, queues, and RSS.
I customized and ran the Pligg-based system rather than writing the underlying engine from scratch. NewsPicker later became Mine Feed.
2001–2008
My first company ran beside my day job.
I built and hosted LAMP-stack sites and web applications for small organizations.
Brio established a pattern I have repeated ever since: do the job in front of me, then keep building at night until I understand the next thing.
1996–1998
Publishing, astronomy, Perl, and Linux before I finished high school.
The surviving record includes a student newspaper byline, a three-person ThinkQuest site, an Internet Guide page, and public troubleshooting threads.
Those projects taught me the pattern I still use: build something real, ask questions in public, and keep going until I understand the system.
What I am doing now
I am VP of Cloud Product Management at VAST Data. I also build Verdify with my adult son and co-founded the Longmont NextWave Initiative with my wife, Emily.