Building at the
Intersection

Full-Stack Developer building AI agents

LangGraphPythonTypeScriptFastAPIRAGQdrantNext.jsReactPostgreSQLDockerLangGraphPythonTypeScriptFastAPIRAGQdrantNext.jsReactPostgreSQLDocker

Featured Work

Agent, RAG and integration projects I built and tested, with the code on GitHub.

Demo recording: Uploading documents to a local knowledge base, then asking questions and watching cited answers stream in

Enterprise RAG Knowledge Base

A cited question-answering API over uploaded documents: FastAPI, Qdrant, hybrid vector plus BM25 search and streamed answers.

hit@1 90% (hybrid, 20 questions)

FastAPILangChainQdrant CloudNext.jsRedisOllama+3

Code

Demo recording: the multi-agent research app running locally

Multi-Agent Market Research Platform

Seven LangGraph agents split a competitive research question into planning, research, analysis, fact checking, report writing and charts, with one human approval step and live progress.

7 agents, 2 parallel stages

LangGraphFastAPINext.js 16Ollama+7

Code

Clio Matter Bridge

A Clio Manage integration with OAuth 2.0, a rate-limit-aware client, a webhook receiver checked against real signed deliveries, and a deny-by-default boundary for anything sent to a language model.

265 assertions, 6 suites, 0 runtime dependencies

Node.jsJavaScriptOAuth 2.0Webhooks+3

Code

API Analytics Hub

Multi-tenant analytics over data synced from a live Shopify store through its Admin GraphQL API into Postgres, with row-level security as the tenant boundary and a mutation harness over the guards.

28 mutation-checked guards

PythonFastAPIPostgreSQLRow-Level Security+5

Code

AreteusML

An ML pipeline for Banking77 intent classification: a fine-tuned ModernBERT model at 91.3% accuracy, served as ONNX INT8 with Dagster, Evidently drift checks and Prometheus and Grafana monitoring.

91.3% Accuracy

PyTorchModernBERTFastAPIONNX Runtime+10

Code

View All Expeditions

Technical Expertise

Skills & Arsenal

Tools I have used on my projects, grouped by area.

AI & Agents

LangGraphLangChainMulti-Agent SystemsTool CallingModel Context ProtocolRAGQdrantLLM IntegrationPrompt EngineeringAgent EvaluationClaude CodeCodexModel Fine-TuningPyTorchHugging Face TransformersONNX Runtime

Voyage Log