Eric Holt

Featured Projects

TraceAgent — Observable Agentic Research Platform

Python · FastAPI · Groq · Tavily · React · PostgreSQL · Docker

A custom agentic pipeline with four sequential stages — planner, search, reflection, and synthesis — where the reflection stage explicitly evaluates whether search results are sufficient and conditionally loops back to search with refined queries before proceeding to report generation. Every internal decision is persisted to the database and streamed to the browser in real time via WebSocket.

Deployed to Hugging Face Spaces as a single Dockerized service · Multi-stage build bundles React/Vite frontend with FastAPI backend · Neon serverless PostgreSQL for persistent run history · Public activity wall with full step traces · Fork any run with a modified query, lineage tracked · Side-by-side comparison of two runs

Live Demo · Code · README

ReconAgent — Autonomous Web Research Agent

Python · LangGraph · FastAPI · Groq · Tavily · React · Docker

A LangGraph agentic pipeline with four sequential nodes — planner, search, reflection, and synthesis — where the reflection node evaluates result adequacy and conditionally loops back to search with refined queries before proceeding to report generation.

Deployed to Hugging Face Spaces as a single Dockerized service · Multi-stage build bundles React/Vite frontend with FastAPI backend · SQLite persistence for research history · GitHub Actions CI/CD

Live Demo · Code · README

AgentFlow – AI-Powered Document Intelligence

Python · LangGraph · FastAPI · Groq · React · Docker

A LangGraph multi-step agent pipeline that classifies uploaded HR documents (resumes, cover letters, interview notes, scorecards) and routes each to a specialized LLM analysis node returning structured JSON with skills, risk flags, and hiring recommendations.

Groq LLM inference + tenacity retry · Vision model support for image/handwritten docs · Deterministic classification pre-check · Single Dockerized service on Render · SQLite session persistence · n8n webhook integration

Live Demo · Code · README

AI Tutor — LangGraph + LoRA (TinyLlama)

Python · LangGraph · LoRA · llama.cpp · Azure

A fine-tuned educational assistant leveraging graph-based reasoning and lightweight quantized inference.

LoRA training pipeline · Graph orchestration · Quantized model serving · Container deployment

Live Demo · Code · README

Transportation-LLM — Local AI Chat via WebGPU

WebGPU · TypeScript · RAG · MLC-AI

An in-browser LLM running client-side inference with retrieval-augmented generation.

WebGPU acceleration · On-device inference · Vector search + prompt augmentation

Live Demo · Code · README

WarehouseBot-RL — Autonomous Picking Simulation

PyTorch · Reinforcement Learning · Q-Learning

A reinforcement learning agent optimizing warehouse path selection through reward-based learning.

Custom environment simulation · Q-learning policy updates · Performance evaluation

Live Demo · Code · README

Get Set Wedding — Production SaaS Planning Platform

React · TypeScript · Supabase · Stripe · Node

A production-grade SaaS platform featuring authenticated dashboards, subscription-based access control, and persistent user state.

Auth/RLS architecture · Stripe billing + plan gating · Refresh-safe sessions · RBAC admin routing · AI tool guardrails

Live App · README

About

I design and build applied AI systems focused on real-world deployment, reliability, and scalable architecture, combining model fine-tuning, agent orchestration, and multi-service infrastructure into modular systems built for production.

My work includes multi-step agent pipelines with conditional looping and reflection, deterministic routing, structured LLM outputs, vision model support, and persistent session state. These are systems that self-evaluate and adapt mid-execution, not just follow a fixed workflow.

I am pursuing a B.S. in Computer Information Systems (Software Programming), deepening my expertise in Python, PyTorch, FastAPI, and distributed AI infrastructure.

Contact

Email · GitHub · LinkedIn