hello, I'm
Butko
I'm a full-stack AI engineerbuilding AI applications, apps and the servers they run on.
Greater Philadelphia
20+
years in enterprise software
12
iOS health apps
656K+
app downloads
2M+
monthly page views
55K+
Docker image pulls
// about
Senior full-stack engineer with two decades of enterprise experience — now building AI-powered applications for Comcast through Wipro.
I work across a platform of 10 products — from conversational AI and natural-language dashboard generation to intent-based routing and enterprise services — and have designed and delivered four full-stack services end to end: requirements, architecture, CI/CD, testing and stakeholder demos.
Before that I spent 20 years in telecom billing with Comverse and Amdocs, deploying RTBS and CCBS systems and helping build the multi-tenant platform that became the Amdocs Digital Brands Suite. I also run my own projects: 12 iOS health apps and websites reaching 2M+ monthly page views.
01
AI & LLM
- Azure OpenAI
- RAG
- Embeddings & semantic search
- AI agents
- MCP
- Local LLM inference
02
Full stack
- Python
- TypeScript
- React
- Next.js
- Node.js
- Electron
- Swift
- Java
03
Data
- PostgreSQL
- Valkey / Redis
- Oracle
- MongoDB
- MySQL
04
Cloud & ops
- Azure
- AWS
- Docker
- Kubernetes
- Terraform
- CI/CD
- Prometheus
- Grafana
05
Integration
- ServiceNow
- REST APIs
- TM Forum APIs
- JBoss Fuse
// selected work
Things I've built
// on the app store
iOS health apps
A family of health trackers with HealthKit, iCloud sync and PDF reports — recently migrated from 100K+ lines of Objective-C to Swift, with 656K+ downloads over the years.
// homelab
Cloud in front, homelab behind
The sites run on cloud servers I manage. The heavy lifting — local LLMs, AI pipelines, image generation and monitoring — runs as Docker services on my own hardware.
production
Cloud
- DigitalOcean droplets
- HestiaCP + nginx
- Next.js on PM2
- MySQL / MariaDB
r&d · at home
Homelab
- Docker services
- llama.cpp local LLMs
- ComfyUI on AMD ROCm
- Prometheus exporters
AI pipelines
Local LLMs with distributed inference power deal scoring and classification for SmartOffs, plus content curation and news aggregation.
Image generation
ComfyUI on AMD GPUs via ROCm — packaged as the public Docker image with 55K+ pulls.
Monitoring
Prometheus metrics from my own exporters for databases and UPS power, watching both the lab and production.
Agent tooling
MCP servers like Subwire that let AI agents talk to every service in the lab.
// open source
Shared with the community
ComfyUI Docker image with AMD ROCm support for GPU-accelerated AI image generation on AMD RX 6000/7000+ GPUs.
A configurable HTTP/REST MCP server for talking to everything in your homelab — and beyond.
A Prometheus exporter for database metrics monitoring.
// contact
Let's talk.
Have a question, an idea or a project in mind? Drop me a line — it lands on a mail server I run myself.





