I take enterprise systems and AI from whiteboard to go-live. 20+ enterprise deployments across healthcare, finance, manufacturing and logistics, plus internal tools and LLM workflows that turn five-hour tasks into under an hour.
This whole site is me.
Keep scrolling: each layer lights up in turn.
I sit between the customer, the product and the code. I listen first, design the shape of the solution, then build the parts configuration can't reach.
enterprise clients delivered across six industries
internal tools adopted across implementation and solution teams
faster on typical tasks: about 5 hours down to under one
years building software, M.S. in Computer Science
The role: Sit with the customer, turn a messy problem into working software, and stay until it is live.
What I bring: Trained through an FDE bootcamp with on-site delivery; 20+ enterprise deployments from discovery to go-live.
The role: Design how AI and existing systems fit together, and scope what is realistic before anything is built.
What I bring: Integration designs across seven HR platforms, scoped before the contract is signed, with AI added where it removes manual work.
The role: Ship LLM features that hold up: retrieval, guardrails, model choice and evaluation.
What I bring: A RAG pipeline with vector search and a knowledge graph, a local-model SQL assistant with lint guardrails, and hands-on model evaluations.
The role: Build agents that take real actions, call tools and know when to hand back to a person.
What I bring: Enterprise agents for a product team, an Instagram DM agent on LangGraph, and a voice assistant that routes on typed confidence.
The role: Explain the product in the customer's language and prove it works with their data.
What I bring: Technical RFP answers, statement-of-work reviews, API proofs in Postman and playbooks other teams reuse.
The role: Take an idea from a blank repo to something people use, front to back.
What I bring: 10+ internal tools in React, Next.js, Python and Swift, adopted across teams, plus my own products on the side.
Two years in enterprise delivery and four in software. I learn each new system by building on it, and I am nowhere near done.
Sit with the customer's HR and IT teams, map their systems and find what the integration has to carry.
Choose the channel (API, file feed, reports), define field mappings and scope custom work before the contract is signed.
Write payloads, pipelines and tools; prove every call in Postman; add AI where it removes manual work.
Reconcile data between systems, run end-to-end tests and stay through cutover until records match.
AI-powered HR tech SaaS · also delivering as a Forward Deployed Engineer
IT services firm
Product startup
Client names stay private. Each group shows the kind of work and the systems involved.
Two-way hiring status sync and historical migration for an academic medical center; reconciled records between two systems for a health system, surfacing 186 mismatches before go-live.
iCIMS · Oracle Taleo · Workday · DayforceEmployee and job-role data sync over OData with SAML-based OAuth, 2,259 job codes mapped to roles, and a bilingual job feed for a mining company.
SAP SuccessFactors · Oracle TaleoEmployee data feeds with data quality rules for a bank, a connector through Azure API Management, historical migrations and automated end-to-end testing.
UKG · WinTeam · PlaywrightFixed field mapping issues, reviewed statements of work and built technical remediation plans to keep accounts healthy.
Workday · SAP SuccessFactorsTools from client work, described by what they do, and things I build on my own time.
Hover for live demoLive demoTeams upload customer files, map columns by hand or with Gemini suggestions ranked by confidence, validate and export clean data.
Hover for live demoLive demoWatches for new documents, extracts structured knowledge with Claude and indexes it as linked notes, a knowledge graph and vector search.
Hover for live demoLive demoDiffs business process versions and checks each step against supported rules. Cut change analysis from hours, sometimes days, to minutes.
Hover for live demoLive demoA Chrome side panel for an analytics SQL editor: eleven lint rules, a schema browser and a chat assistant that writes read-only queries with a local model.
Hover for live demoLive demoA system-wide macOS utility in Swift that fixes each typed word from sentence context with an on-device model, only when it's confident.
Hover for live demoLive demoA push-to-talk voice assistant for macOS that scores intent and risk with typed confidence, then answers with local or cloud models.
Hover for live demoLive demoAn agent in Instagram DMs that turns shared reels into searchable summaries and tags, then finds them again from plain-language asks.
Hover for live demoLive demoA creator-first recipe platform for chefs, with authoring, collections, reviews and creator analytics.
Ran local evaluations of a full-duplex speech-to-speech model.
Set up a multi-provider model gateway in Docker.
Pick Ollama and llama.cpp or Claude and Gemini on privacy and cost, so the SQL and typing assistants run locally by default.
Pick a skill to see how deep it goes and where I've used it, or pick a project in the middle to see the skills behind it. Every line links a skill to work that proves it.