01MLOps · Time series
ML demand forecasting in production
Over 100,000 SKU+Store combinations, forecast weekly by a routing policy that picks a model per cluster.
- Series, weekly
- 100k+
- Accuracy by cluster
- 70-82%
Selected work
Five systems I designed, built and operate. Four run at a retail chain in Colombia. The fifth is my own infrastructure. Every one has a working demo you can click.
Selected work
Five systems I designed, built and operate. Four run at a retail chain in Colombia. The fifth is my own infrastructure. Every one has a working demo you can click.
02Data platform · Generative AI
Market price intelligence platform
Weekly public data, parsed into a 52-week forecast the buying team takes into supplier negotiations.
- Forecast horizon
- 52 wk
- Products covered
- 100s
Buy now
03Applied AI · Guardrails
AI assistant for business data
A Telegram assistant for questions about sales, inventory and purchasing. Every figure it gives comes from a query it actually ran.
- Numbers the model makes up
- Zero
- Embedded/Vector DB
- Golden queries
04Full stack · Internal product
Stock rebalancing engine
Analysts used to work out every inter-store transfer by hand. The engine proposes what moves where, and they run it themselves.
- Per transfer plan
- 4h → 5min
- Run by the analysts
- Self-serve
05Multi-agent · Local-first
Personal AI operating system
The memory is markdown in Git, so it outlives the runtime that reads it.
- Before anything it can't undo
- Asks first
- Git is the source of truth
- Markdown
All five, at a glance
01MLOps · Time series
ML demand forecasting in production
- Series, weekly
- 100k+
- Accuracy by cluster
- 70-82%
Over 100,000 SKU+Store combinations, forecast weekly by a routing policy that picks a model per cluster.
View full case study: ML demand forecasting in production02Data platform · Generative AI
Market price intelligence platform
- Forecast horizon
- 52 wk
- Products covered
- 100s
Weekly public data, parsed into a 52-week forecast the buying team takes into supplier negotiations.
View full case study: Market price intelligence platform03Applied AI · Guardrails
AI assistant for business data
- Numbers the model makes up
- Zero
- Embedded/Vector DB
- Golden queries
A Telegram assistant for questions about sales, inventory and purchasing. Every figure it gives comes from a query it actually ran.
View full case study: AI assistant for business data04Full stack · Internal product
Stock rebalancing engine
- Per transfer plan
- 4h → 5min
- Run by the analysts
- Self-serve
Analysts used to work out every inter-store transfer by hand. The engine proposes what moves where, and they run it themselves.
View full case study: Stock rebalancing engine05Multi-agent · Local-first
Personal AI operating system
- Before anything it can't undo
- Asks first
- Git is the source of truth
- Markdown
The memory is markdown in Git, so it outlives the runtime that reads it.
View full case study: Personal AI operating system- 01MLOps · Time series ML demand forecasting in production 100k+ Series, weekly Case study
- 02Data platform · Generative AI Market price intelligence platform 52 wk Forecast horizon Case study
- 03Applied AI · Guardrails AI assistant for business data Zero Numbers the model makes up Case study
- 04Full stack · Internal product Stock rebalancing engine 4h → 5min Per transfer plan Case study
- 05Multi-agent · Local-first Personal AI operating system Asks first Before anything it can't undo Case study