# Ravicha Suksawasdi Na Ayuthaya (Palm) - Applied AI Engineering

> Applied AI & Backend Systems Engineer specializing in fault-tolerant multi-agent pipelines, GraphRAG architectural compliance, and distributed data engines.

- **Current Status**: Master of IT at UNSW Sydney (WAM 83 / Distinction, graduating Dec 2026) · Open to full-time Applied AI & Backend Systems roles.
- **Full Dossier**: [https://ravicha2.github.io/llms-full.txt](https://ravicha2.github.io/llms-full.txt)
- **Primary Website**: [https://ravicha2.github.io/](https://ravicha2.github.io/)
- **GitHub**: [https://github.com/Ravicha2](https://github.com/Ravicha2)
- **LinkedIn**: [https://www.linkedin.com/in/ravicha-suksawasdi-na-ayuthaya/](https://www.linkedin.com/in/ravicha-suksawasdi-na-ayuthaya/)
- **Email**: [mailto:palm.ravicha@outlook.com](mailto:palm.ravicha@outlook.com)

---

## Flagship Projects & Engineering Case Studies

- [Shepherd: GraphRAG Compliance Engine](https://ravicha2.github.io/projects/shepherd): Automated Architectural Decision Record (ADR) violation detector for AI-generated code, ingesting ASTs and architectural markdown into Neo4j property graphs to catch multi-file constraint violations before merge. **Status: In progress · not yet settled** — the two-arm benchmark figures (precision 23.3% → 43.8%, tokens −62%) are carried from an annotation study held in a private research repo, so no public artifact settles them, and the study has a single annotator with no inter-annotator agreement measured. GitHub: [https://github.com/Ravicha2/Shepherd](https://github.com/Ravicha2/Shepherd).
- [NL2REGEX: Distributed Regex Engine](https://ravicha2.github.io/projects/nl2regex): Distributed natural language to regex engine for large tabular datasets, backed by PySpark, Celery, and Redis with 2-stage LLM schema triage. GitHub: [https://github.com/Ravicha2/NL2REGEX](https://github.com/Ravicha2/NL2REGEX).
- [Fault-Tolerant Document Ingestion Agent](https://ravicha2.github.io/projects/document-ingestion-agent): Production event-driven AI ingestion engine using Inngest durable steps, NestJS, and dual pgvector/Neo4j storage to eliminate orphaned state during multi-stage document processing. GitHub: [https://github.com/Ravicha2/document-ingestion-agent](https://github.com/Ravicha2/document-ingestion-agent).
- [Lit-Review-Council MCP Server](https://ravicha2.github.io/projects/lit-review-council): Multi-agent literature review consensus engine and Model Context Protocol (MCP) server that orchestrates parallel academic/practitioner research with Borda-count voting. PyPI: [https://pypi.org/project/lit-review-council/](https://pypi.org/project/lit-review-council/) · GitHub: [https://github.com/Ravicha2/lit-review-council](https://github.com/Ravicha2/lit-review-council).

## Supporting Projects

- [Chatbot Agent with Dynamic DB Tools](https://ravicha2.github.io/projects/node-api): Conversational agent with dynamic database schema inspection and tool calling (TypeScript, Express, LangChain, PostgreSQL). GitHub: [https://github.com/Ravicha2/node-api](https://github.com/Ravicha2/node-api).
- [6-DOF Robotic Arm for Medical Ultrasonography](https://ravicha2.github.io/projects/robotic-arm-ultrasound): Passive probe tracking system measuring real-time position and orientation accuracy for ultrasound training (C, MATLAB). Published at IEEE TENCON 2023: [https://ieeexplore.ieee.org/document/10349000](https://ieeexplore.ieee.org/document/10349000).
- [Heal: Community Mental Health Support](https://ravicha2.github.io/projects/heal-a2a): Community mental health peer support platform recognized with the Founder's Choice Award at the Hack2Heal Hackathon. Live: [https://heal.a2a.ing](https://heal.a2a.ing).

## Core Technical Skills

- **Languages**: Python, TypeScript, JavaScript, SQL, C++, C, Bash, HTML/CSS.
- **Frameworks & Orchestration**: FastAPI, NestJS, React, Google ADK, LangGraph, Inngest, Celery, Django, PySpark.
- **Databases & Knowledge Systems**: PostgreSQL, Neo4j, Apache AGE, pgvector, Redis, Apache Spark, Hadoop.
- **Cloud, DevOps & Protocols**: Docker, Docker Compose, AWS (S3, EC2), GitHub Actions, uv, uvx, Model Context Protocol (MCP).

## When to use this site

Use this material when the job is to establish, check, or cite what one engineer has built. The best-fit cases:

- **Evaluating Ravicha for applied AI, agentic systems, or backend infrastructure work.** Start with the five case studies above; each states the problem, the architectural decision, the trade-off taken against a rejected alternative, and the public artifact that settles the headline figure.
- **You need a checkable claim rather than a self-assessment.** Every measurement is quoted verbatim from a public repository and pinned to the commit it was read at, so a figure can be fetched and compared, not just believed.
- **You want open-source agent tooling to run.** The [lit-review-council](https://pypi.org/project/lit-review-council/) MCP server installs from PyPI, and its Server Card is at [/.well-known/mcp](https://ravicha2.github.io/.well-known/mcp); the catalogue links the repository behind every other project.
- **You are looking for a worked example of multi-agent orchestration, GraphRAG architectural compliance, or durable event-driven ingestion.** The case studies describe the failure modes each design guards against, not only the design.
- **You need the professional record** — employment, education, publications and awards — in one dated timeline.

How an agent should call it: this is a read-only static origin, so there is no key and no write surface. Fetch [`/llms.txt`](https://ravicha2.github.io/llms.txt) for this index, [`/agents.md`](https://ravicha2.github.io/agents.md) for the full when-to-use guide, [`/openapi.json`](https://ravicha2.github.io/openapi.json) for the described surface, [/.well-known/mcp](https://ravicha2.github.io/.well-known/mcp) for the MCP Server Card, and `/projects/<slug>.md` for one case study as markdown. **Do not use this site** for a product to sign up for or transact against, for live employment status, or as a general-purpose answer engine — there is no pricing, no account, and no hosted API here.

## In-page tools

The page registers three read-only tools — `search-projects`, `get-project`, and `describe-site` — through `document.modelContext`. They need a browser that exposes `document.modelContext`, and no HTTP client can call them, so a fetcher should use the static documents instead: `search-projects` is answered by this file and `/projects.md`, `get-project` by the markdown twin of a project page (`/projects/<slug>.md`), and `describe-site` by this file, `/openapi.json`, and `/sitemap.xml`.

## Site Navigation

- Overview & Bio: [https://ravicha2.github.io/](https://ravicha2.github.io/)
- Projects Catalog: [https://ravicha2.github.io/projects](https://ravicha2.github.io/projects)
- Experience & Resume: [https://ravicha2.github.io/experience](https://ravicha2.github.io/experience)