CometLabsBuilding practical AI systems.
AI Agent · RAG · Automation · ERP · Product EngineeringCometLabs pulls documents, tools, and workflows into one focused AI operating layer, solving real operational inefficiencies with practical AI systems.
Featured Projects
Selected AI products and experiments.
Agentic Workflow Automation Platform
Turn plain-language requests into repeatable business workflows
Promotion Portal
A promotion reward automation management system
Capabilities
What I Build
AI Agents
Agent systems that plan and execute work autonomously while controlling legacy APIs and tools.
RAG Systems
Enterprise retrieval platforms with hybrid embedding search and source verification.
Automation Tools
Automation tools that unify repetitive file processing, parsing, and report delivery.
AI-powered ERP
Intelligent ERP experiences that replace complex menu navigation with natural language workflows.
AI Radar
Tracking AI shifts & technical applications.
Codex
OpenAI's agentic coding CLI that works directly against a repository: it reads code, implements changes, reviews diffs, and verifies its own work with tests from the terminal.
Claude Code
Anthropic's terminal-based agentic coding assistant, running the newest Claude models. It reads and edits project files, runs builds and tests, and can spawn specialized subagents for parallel work.
MCP
Model Context Protocol is an open standard that lets LLMs and AI agents communicate with external data sources and developer tools such as databases, filesystems, Slack, and GitHub.
History
Project Timeline
Agentic Workflows & Multi-Agent Systems
Researching and building practical automation products with AI agents and MCP integrations.
- Built an agentic workflow automation platform that turns plain-language requests into repeatable business workflows.
- Orchestrated a hierarchical multi-agent dev team across coding CLIs with automated QA and risk audits.
- Maintained an experiment-driven AI development archive with daily commits.
RAG & Document Automation / AI Service Experiments
Developed and tested an internal document analyzer based on retrieval-augmented generation.
- Built a high-precision RAG pipeline with hybrid keyword and vector search.
- Implemented source matching and integrated viewer highlighting for AI Document Assistant.
- Created small business automation scripts and compared performance.
Business Automation & ERP Improvement
Converted manual Excel workflows into database-backed systems with automated approvals.
- Integrated promotion reward calculation with multi-stage approval in Promotion Portal.
- Reduced operational mistakes through automated Excel parsing and validation.
- Validated a reduction in settlement time from three days to under one hour.
Engineering Logs
Latest Build Notes
Building an AI Dev Team: Hierarchical Multi-Agent Orchestration
How I structured two coding CLIs and a set of model-tiered workers into a role-based agent company with a team lead, a manager, and specialized implementers.
Self-Generating Chatbot QA: An LLM-as-Judge Evaluation Pipeline
An automated pipeline that generates test questions from a chatbot's own source documents, scores answers with a frontier LLM, and leaves final judgment to a human.
Automated Pre-Deployment Risk Audits with a Read-Only Code Agent
A recurring feature-ledger review where a read-only code agent deep-reads candidate features and emits strict structured risk reports before deployment.
Let's build practical AI systems.
I welcome collaboration around AI agents, intelligent products, and practical business automation systems.