Data & AI Systems Engineer · Washington D.C.

Building systems
that make organizations
measurably faster

I design retrieval pipelines, automation workflows, and knowledge infrastructure — deployed in production, not just prototyped. Currently building Savant and leading data architecture & automation at NOAA/NWS.

80% Reporting cycle reduction @ NWS

RAG Production hybrid BM25 + FAISS

2 Active deployments

View Projects LinkedIn →

scroll

Selected Work

Projects & Case Studies

01 / Production RAG System

Savant — Knowledge Retrieval Infrastructure

Built a production-grade RAG system implementing Anthropic's Contextual Retrieval architecture. Hybrid BM25 + FAISS search with Reciprocal Rank Fusion consistently outperforms single-method retrieval on ambiguous queries. Section-aware chunking for PDF and DOCX, GPT-4o-mini contextualization pipeline, multi-threaded embedding, and S3-backed index with automatic rebuild. Admin analytics dashboard tracks fallback rates, query patterns, and source attribution.

BM25 + FAISS Contextual Embeddings RRF Reranking Python Anthropic API S3 Gov Contractor

Building In Development

02 / Federal Data & Automation

NWS Logistics Workflow Automation

Redesigned fiscal and logistics reporting workflows for the NWS Logistics Management Branch — infrastructure supporting NWS, FAA, and DOD operations. Rebuilt legacy reporting infrastructure using Oracle SQL scripting and workflow automation, replacing manual multi-day reporting cycles with automated, auditable infrastructure. Also built AI-assisted engineering workflows using custom Gemini agents to accelerate SQL generation and system architecture decisions.

80% Cycle Reduction Oracle SQL SQL Scripting Gemini Agents NWS · FAA · DOD

80% Cycle Reduction

03 / Custom GPT Deployment

Potencia GPT — Nonprofit AI Infrastructure

Designed and deployed a custom GPT system for Potencia, an immigrant-focused nonprofit. Built organizational knowledge into a GPT assistant to support internal operations, donor communication, and program tracking. Delivered with prompt library, usage documentation, and staff training. Currently serving as Advisory Board member providing ongoing AI strategy.

GPT Deployment Prompt Engineering Org Knowledge Systems Nonprofit Advisory Board

Live In Production

Technical Stack

What I Build With

Retrieval / AI

BM25 Hybrid Search
FAISS Vector Store
Contextual Embeddings
RRF Reranking
Anthropic Claude API
OpenAI GPT-4o

Data / Backend

Python
Oracle SQL
Relational Modeling
AWS S3
ThreadPoolExecutor
REST APIs

Systems / Infra

Workflow Automation
Document Processing
PDF / DOCX Chunking
Admin Dashboards
Query Analytics
Agent Pipelines

Domain

Federal Gov Contracting
NWS / FAA / DOD
Enterprise RAG
Fiscal Automation
Knowledge Systems
AI Strategy

Background

About Alex

Alex Samadi

Alex Samadi

Data & AI Systems Engineer · Washington D.C.

I'm a Data & AI systems engineer based in Washington D.C., currently building Savant — a knowledge retrieval system targeting government contractors — while leading data architecture & workflow automation at NOAA/NWS via Innovim.

My work sits at the intersection of data architecture and AI infrastructure. At NWS, I rebuilt fiscal reporting systems from the ground up using Oracle SQL and relational modeling. At Savant, I implemented Anthropic's full Contextual Retrieval stack in production — BM25 hybrid search, contextual embeddings, section-aware chunking, and RRF.

I'm most interested in hard operational problems at the intersection of AI, systems design, and real organizational impact — federal, startup, or my own company.

Let's Work Together

Have a hard
problem to solve?

alexsamadi321@gmail.com linkedin.com/in/alex-samadi