Enterprise architecture · Cloud · AI
Architecting intelligent systems for the real world.
I’m Abiola Babsalaam, an enterprise architect. This is where I document my open-source projects, local AI experiments, and what I learn while building them.
Building in publicOpen-source projectsEnterprise + edge
SYSTEM MAP / 2026
Intelligence, by design.
Areas I work in
Enterprise architecture, cloud infrastructure, and practical AI.
Enterprise architecture
Platforms, technical decisions, operating constraints, and roadmaps.
Cloud & AI infrastructure
Secure foundations for modern workloads, generative AI, and agentic systems.
Open-source experiments
Local and edge AI prototypes that people can inspect, reproduce, and extend.
Projects
Open-source work.
Projects and experiments involving local AI, edge hardware, model evaluation, and developer tooling.
Featured build · Local AI
Teacher in a Box
A local classroom that diagnoses, teaches, trains, and evaluates small language models.
I built it to make fine-tuning easier to inspect. It diagnoses weak skills, lets a person review targeted lessons, trains locally, and compares the result on a held-out exam. Lemonade provides the local inference layer.
Model tooling · JavaScript
FineTune Studio
A unified local desktop trainer for text, audio, video, and Liquid models, with a hosted Liquid YAML builder.
Explore repository ↗Edge research · Python
EdgeSLM Text-to-SQL
A reproducibility package for evaluating small open-weight text-to-SQL models on an 8 GB edge device.
Explore repository ↗Computer vision · Python
Edge Litter Detection
Real-time detection, tracking, counting, and SQLite evidence for cleanup and recycling workflows.
Explore repository ↗Working principles
Technology should be ambitious and operationally grounded.
I care about the complete path from a promising idea to a system someone can understand, operate, and improve.
Clarity before complexity
Make decisions, dependencies, and tradeoffs visible before adding technology.
Evidence over spectacle
Document baselines, constraints, and results so progress can be reproduced.
Local where it matters
Keep sensitive data and model interactions close when privacy and control matter.
Open by default
Share practical code and honest documentation that help the next builder move faster.
Blog
Notes from what I’m building.
Longer posts about local AI, evaluation, infrastructure, and the lessons that only show up once I start building.
01
Building Teacher in a Box: a local classroom for small models
A look at why I separated diagnosis, curriculum review, training, and evaluation, plus where Lemonade fits.
About
Architecture is how bold ideas become dependable systems.
I work at the intersection of enterprise architecture, cloud computing, and AI infrastructure. I turn complex technical and operational requirements into secure, scalable systems with clear business value.
My passion is making advanced AI more practical. Through open-source work, I explore generative and agentic AI, local and edge systems, financial services, and the evidence required to move from prototype to trustworthy solution.