selected work
Problems, and what changed after
A few engagements I can talk about. Each one has the real problem, how I approached it, and what actually moved. Client names are withheld. The numbers and methods are not.
Cut a SaaS startup's AWS bill by ~42%
A Series-A SaaS team's AWS bill had tripled in a year and nobody could say why. A focused FinOps pass cut it ~42% without touching reliability.
Recovered a production cluster from a CrashLoopBackOff outage
A node upgrade left an entire production namespace in CrashLoopBackOff. Mitigated in under an hour, root-caused to a probe and config-map mismatch, and fixed so it can't recur.
Replaced an internet-facing VPN appliance with a zero-trust mesh
A team running an internet-facing VPN appliance, the exact category behind a wave of 2024 CVEs, moved to a Tailscale and Cloudflare Tunnel mesh and removed the public concentrator entirely.
Shipped a self-hosted RAG assistant with an LLM gateway
A team with a promising RAG demo couldn't ship it. Accuracy was unmeasured and API costs were unpredictable. A measured pipeline behind an LLM gateway with hybrid serving made it production-ready and ~70% cheaper.
Zero-downtime migration to Kubernetes with multi-cloud ingress
A team moving from hand-managed VMs to Kubernetes needed it done without an outage. A staged, GitOps-driven migration with weighted ingress shifted traffic gradually and reversibly, with zero downtime.
Built a customer-support AI agent that deflected ~60% of tickets
A support team was drowning in repetitive tickets that the help docs already answered. An AI agent over those docs and past tickets, with order-lookup tools and a clean human handoff, took ~60% of volume off the queue.
Built a RAG assistant over 12,000 internal docs with citations
Employees wasted hours hunting through 12,000 scattered internal documents for answers that existed somewhere. A RAG assistant with inline citations turned that into a one-line question with a sourced answer.
Cut a company's LLM API bill by ~65% with a gateway
A product team was sending every request to a frontier model and watching the bill climb past usefulness. A gateway with caching, model routing, and prompt trimming cut spend ~65% with no drop in output quality.
Took a support bot from 72% to 96% answer accuracy
A support bot was already live and quietly wrong about a quarter of the time, with no one measuring it. An eval harness, fixed retrieval and grounding, and guardrails took accuracy from 72% to 96%.
Automated a manual ops workflow and saved ~20 hours a week
An ops team hand-sorted every incoming request, reading, categorizing, and routing each one by hand. An LLM pipeline that triages and routes with a human review step gave back ~20 hours a week.
Need a result like one of these?
Book a free 30-minute call. We diagnose it together, and you walk away with a plan you can act on. You’ll get a straight read either way.