blackroad-os/alexa-amundson-resume
9
0
Fork 0
mirror of https://github.com/blackboxprogramming/alexa-amundson-resume.git synced 2026-03-18 05:34:08 -05:00
Code Issues Packages Projects Releases Wiki Activity

kpi: auto-update metrics 2026-03-13

Browse Source 
RoadChain-SHA2048: c645c1292ab1555e
RoadChain-Identity: alexa@sovereign
RoadChain-Full: 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
This commit is contained in:
Alexa Amundson
2026-03-13 23:16:12 -05:00
parent 0c714c106c
commit ec7b1445b5
25 changed files with 815 additions and 1112 deletions
Download Patch File Download Diff File Expand all files Collapse all files

66
roles/02-ai-ml-engineer.md
View File

@@ -8,64 +8,44 @@ amundsonalexa@gmail.com | [github.com/blackboxprogramming](https://github.com/bl
## Summary
AI/ML engineer operating 27 deployed language models (48.1 GB) across a distributed edge fleet with 52 TOPS of dedicated AI acceleration. Builds custom model pipelines, inference APIs, and AI-powered automation across a 7-node infrastructure serving 48+ domains.
Cloud AI APIs are expensive and you don't own the data. Deployed 27 language models on-premise across edge hardware with 52 TOPS of dedicated acceleration — full inference sovereignty at a fraction of the cost.
---
## Experience
### BlackRoad OS | Founder & AI Lead | 2025–Present
### BlackRoad OS | Founder & AI/ML Engineer | 2025–Present
**Model Deployment & Operations**
- Deploy and manage 27 Ollama models (48.1 GB) across 3 edge nodes
- Fine-tuned 4 custom CECE personality models for domain-specific generation
- Operate 2× Hailo-8 NPUs (26 TOPS each = 52 TOPS total) for accelerated inference
- Built Ollama Bridge SSE proxy for streaming model responses to web clients
**The Problem: AI Without Vendor Lock-In**
- Needed persistent, private AI inference without per-token API costs or data leaving the network
- Deployed 27 Ollama models (48.1 GB) across 3 Pi 5 nodes — installed 2x Hailo-8 NPUs (52 TOPS total) for hardware acceleration
- Fine-tuned 4 custom CECE personality models for domain-specific generation — models that don't exist anywhere else
**AI Infrastructure**
- Designed multi-node inference routing across Pi 5 fleet (20 GB RAM total)
- Built AI image generation hub (images.blackroad.io) with 4 backend agents (DALL-E, Flux, SDXL)
- Manage model lifecycle: deployment, monitoring, memory optimization, thermal throttle prevention
- Reduced inference node temperature from 73.8°C to 57.9°C by isolating runaway generation loops
**The Challenge: Thermals Kill Edge AI**
- Inference on $80 hardware generates heat. A runaway generation loop pushed one node to 73.8°C — approaching thermal shutdown
- Built power monitoring (cron every 5 min), CPU governor tuning, and voltage optimization — stabilized fleet at 42°C average
- Reduced GPU memory allocation from 256MB to 16MB on headless nodes, capped frequencies, applied conservative governors — no inference quality loss
**APIs & Integration**
- Built CECE API (FastAPI) for custom model interaction and TTS generation
- Created code search engine indexing 354 repos with FTS5 full-text search
- Operate 11 PostgreSQL databases for AI application state
- Deploy AI services via Cloudflare Workers (99 Pages, 22 D1 databases)
**Data & Knowledge Systems**
- Manage 230 SQLite databases (1.4 GB) for agent memory, metrics, and state
- Built FTS5 knowledge index across 156,675 memory entries
- 111 registered systems tracked in systems database
- Custom RoadC programming language with tree-walking interpreter
**Automation & Monitoring**
- Daily automated KPI collection tracking 60+ AI and infrastructure metrics
- Power optimization across all AI nodes (CPU governors, voltage tuning)
- Self-healing autonomy scripts ensuring model availability
**The Stack: From Model to API to User**
- Built Ollama Bridge SSE proxy for streaming model responses to web clients in real-time
- AI image generation hub with 4 backend agents (DALL-E, Flux, SDXL, FAL) — single API, best-model routing
- FTS5 knowledge index across 156,675 memory entries — models can search their own history across 230 SQLite databases
---
## Technical Skills
**AI/ML:** Ollama, Hailo-8 NPU, DALL-E, Flux, SDXL, custom fine-tuning, LLM orchestration
**Languages:** Python (470 repos), JavaScript (114 repos), TypeScript (85 repos), Bash, Go, C
**Frameworks:** FastAPI, Next.js, React, Node.js
**Infrastructure:** Raspberry Pi 5 fleet, Docker Swarm, WireGuard, Cloudflare
**Databases:** PostgreSQL, SQLite/FTS5, Cloudflare D1, KV stores, Qdrant (vector)
**Tools:** Ollama, Docker, GitHub Actions, Gitea, Wrangler
Ollama, Hailo-8 NPU, DALL-E, Flux, SDXL, FastAPI, Python, FTS5, Docker
---
## Metrics
| Metric | Value |
|--------|-------|
| Models deployed | 27 (48.1 GB) |
| AI acceleration | 52 TOPS |
| Custom models | 4 fine-tuned |
| Image gen agents | 4 |
| Databases | 241 (11 PG + 230 SQLite) |
| Lines of code | 7,212,320 |
| Repos | 1,810 |
| Metric | Value | Source |
|--------|-------|--------|
| AI Models | *live* | services.sh — ollama list via SSH |
| Model Size (GB) | *live* | services.sh — ollama list via SSH |
| Lines of Code | *live* | loc.sh — cloc + fleet SSH |
| Total Repos | *live* | github-all-orgs.sh — gh api repos (17 owners) |
| SQLite DBs | *live* | local.sh — find ~/.blackroad -name *.db |
| Docker Containers | *live* | services.sh — docker ps via SSH |