I build AI systems.
I focus on machine learning foundations, backend systems, and performance-aware design.
I care about how things work, why they fail, and how to make them reliable.
Currently preparing for Google Summer of Code 2026 and long-term open-source contribution.
Designing LLM-based agent systems with structured reasoning, clear interfaces, and deployable backends.
Implementing models from first principles to understand training dynamics, gradients, and failure modes.
Using Python, SQL, and backend tools to turn data into decisions, not charts.
Multi-Agent Orchestration Framework
FastAPI · Docker · LLMs
Backend-first agent architecture with task routing, tool usage, and reproducible deployment.
Neural Network From Scratch
NumPy
Forward pass, backpropagation, optimization, and numerical stability.
Data Analysis Projects
Python · SQL
Clean EDA focused on logic and insight.
Studying non-determinism and hallucination in LLMs.
Understanding why identical prompts produce different outcomes, and what that means for real systems.
The Calculus of Cognition
Neurons → ANN → Backpropagation
Award-winning technical seminar.
Mathematics of Data
Signals, images, tensors — explained clearly.
- Clean repositories
- Clear READMEs
- Reproducible environments
- Gradual testing and CI
- Code meant to be read by others
I’m interested in organizations working on:
- Scientific computing and ML tooling
- Python ecosystems with performance-critical components
- AI infrastructure and backend systems
I value depth over speed, and long-term contribution over short-term visibility.
📧 [email protected]
🐦 https://x.com/GiGiKoneti
💻 https://github.com/GiGiKoneti
Simple. Clear. Built to last.