
LLM & Agents
Argus
A wrapper for the raw Anthropic SDK that logs every tool call, cost, and error, with a live dashboard for multi turn agent sessions. One caching fix cut token cost 86%.
- Next.js
- FastAPI
- PostgreSQL
- Redis
- SSE
- Docker
Full-Stack & AI Engineer
I'm a full-stack engineer who ships production AI and ML end to end, from the data and model layer to the interface people actually use. I go deep in the domain I'm building for and take most of it from idea to deployment on my own.

Selected work
Every one is live and open source. Click through to try it or read the code.

LLM & Agents
A wrapper for the raw Anthropic SDK that logs every tool call, cost, and error, with a live dashboard for multi turn agent sessions. One caching fix cut token cost 86%.
EnergyML & Data
Province wide flare and vent intelligence over Petrinex data. Scores anomalies against AER Directive 060 with an isolation forest and maps every operator in Alberta.

Full-Stack
Upload a CSV and ask questions in plain English. Lumin answers with auto generated charts, flags anomalies in numeric columns, and exports the findings.
EnergyLLM & Agents
Ask questions across 8 AER directives and 200 wells. Hybrid retrieval with a cross encoder reranker grounds every answer in cited source passages.
EnergyOptimization
Picks which inactive wells an operator should close to hit its regulatory quota at lowest cost, using constraint optimization against a live budget slider.

Full-Stack
An AI crop advisory built for Zimbabwean smallholder farmers, turning local growing conditions into plain language guidance a farmer can act on.
About
I'm Anesu, a full-stack engineer who builds AI and ML, and I finished a computer science degree at Cleveland State University in 2026. I approach building the way a founder would: start from a real problem in a specific domain, learn that domain properly, then ship the smallest real thing that actually solves it.
Most of what I build sits on the raw Anthropic SDK with no framework in between, because I want to know exactly what the model is doing and what each call costs. I tend to build the whole stack myself: the data and retrieval underneath, the model and agent logic in the middle, and the interface someone actually clicks. Most of what I build I take from idea to production on my own, and it all ships live.
Lately I've been going deeper on the machine learning side: anomaly detection over messy public data, constraint optimization, and tightening retrieval so answers stay grounded in real sources. The part I care about most is where a model stops being a demo and has to hold up under real use.
Based in the US and open to roles across the US and Canada, remote or on site.
What I work with
Languages
AI & LLM
Machine Learning
Full-Stack
Data & Infra
Contact
Whether it's a role, a project, or a question, my inbox is open. Email is the fastest way to reach me.