Projects
Broader project range across AI systems, data workflows, automation, and technical problem solving. For deeper technical detail, see the case studies.
Flagship Projects
Analytics Agents
DataInsight
AI-powered analytics platform for non-technical users, combining natural language requests, agent-based reasoning, automated preprocessing, and structured analysis workflows.
- FastAPI, Streamlit, OpenAI APIs, and modular processors
- Cleaning, planning, analysis, and response generation pipeline
- Clarification loops and adaptive error recovery for messy data
Agent Systems
Alden
Local-first planning assistant designed around bounded autonomy, tool routing, structured memory, and execution reliability.
- Multi-model reasoning and tool orchestration
- Stateful architecture with SQL-backed memory
- Clear human approval boundaries
Automation and Agents
Budget Agent
Personal finance agent that accepts Telegram messages, receipts, and transaction files, then turns them into structured budget records, analysis, and forecasting.
- OCR and CSV transaction ingestion
- SQLite-backed persistence with specialist-agent routing
- Repayment planning projected to save about $100 per month over 4 years
Personal Assistant Systems
MIRA
OpenClaw-style assistant system with a custom IO hub, Telegram interface, routing layer, scheduler, and persistent memory designed for real assistant workflows.
- FastAPI and Telegram-based communication architecture
- Custom dispatcher, router, and scheduling components
- Service-oriented assistant design with persistent memory
Supporting Work
Compass
Repository-aware maintenance agent that audits repositories, proposes documentation and portfolio improvements, and supports approval-first PR workflows.
LifeNotes
Note intelligence project exploring capture, retrieval, and workflow-aware information reuse.
Aircraft Inventory Simulation
Discrete-event simulation project modeling aircraft parts inventory behavior, capacity constraints, and long-horizon operational tradeoffs.
RL and Architecture Experiments
Smaller exploratory work around policy optimization, constrained environments, routing patterns, and orchestration reliability.
Data Analysis and Visualization
Applied analysis work using Python and statistics to clean data, inspect patterns, and communicate decisions more clearly.
Additional Work
- Applied AI systems across different problem types
- Engineering work that connects models, tools, and workflows
- Projects shaped by constraints, iteration, and implementation detail