Sam Griffith
01 — About

Engineering elegant platforms. Harnessing AI. Keeping tech human.

I'm a software engineer and tech lead with a path into the craft that didn't start with code. I spent the first decade of my career in financial operations and business analysis before transitioning into engineering, and that background is load-bearing. Understanding how systems serve people before learning how to build them shaped everything about how I approach the work — and it's the direct reason I moved into a technical lead role quickly once I made the switch.

I approach this craft the way a good tradesperson approaches their work: complexity is a tool, not a default. I reach for it only when the problem actually demands it. The measure of a well-built system isn't how impressive it looks on the day it ships — it's whether the engineer who inherits it two years later can understand it, trust it, and extend it without fear. That's the standard I build to.

My day-to-day lives at the intersection of Node.js, TypeScript, and Java services stitched together with event-driven patterns on AWS and Kubernetes. Outside of production work I'm deep in agentic AI — building personal projects with LangGraph and CrewAI and driving adoption of AI-augmented workflows across my team. I care as much about the people using what I build as the engineers maintaining it. Technology that doesn't genuinely serve both groups isn't finished yet.

When I'm not building production systems, I'm exploring the edges of what agentic AI workflows can actually do at scale — which is less "impressive demo" territory and more "will this hold up at 3am on a Tuesday" territory. I'm based in the northern suburbs of Cincinnati, OH and selectively open to remote opportunities where the problem space is genuinely interesting or the mission of the organization is compelling and meaningful.

02 — Experience

Where I've built things

Fidelity Investments

Senior Software Engineer / US Technical Lead

Oct 2022 — Present

Leading the US engineering side of a cloud-native communications platform that handles tens of millions of financial, regulatory, and marketing communications annually on AWS and Kubernetes. The work is a long-horizon migration — decomposing legacy document generation systems into event-driven microservices with the kind of observability and reliability standards a regulated environment demands. Alongside that, I've been an active advocate for AI-augmented development across the team — creating and distributing GitHub Copilot hub-and-spoke workflows via a shared Git repository, running adoption sessions, and helping colleagues build fluency with AI tooling as a natural part of how we write and ship software.

Fidelity Investments

Senior Business Analyst — Camunda & Workflow Systems

Aug 2020 — Oct 2022

Workflow automation for a communications product area — identifying automation candidates, writing specifications that drove engineering execution, and serving as the primary bridge between business stakeholders and the engineering squad. This is where the pattern that defines how I work now got established: I already understood how systems serve people before I understood how to build them, and that sequence mattered. It's the direct reason I was able to move into a technical lead role quickly after transitioning into engineering.

Fidelity Investments

Brokerage Operations & Client Services

2012 — 2020

Eight years in client-facing and operational roles across brokerage and financial services. The compliance awareness, stakeholder judgment, and process thinking I developed here aren't soft skills that sit adjacent to my engineering work — they're load-bearing. Secure-by-default design, governance-oriented architecture, and the instinct to ask who a system actually serves before asking how it works all trace back to this foundation.

03 — Projects

Things built outside of work

Newsfloor Agent

Python LangGraph CrewAI AWS Bedrock

A personalized AI research digest, built and operated as a working agentic system. Each day it autonomously sources, filters, and synthesizes content across 20 topic areas — multi-agent orchestration, observability, governance, reliability, and more — tailored to where I'm focusing my work. As much a learning tool as a demonstration: running it surfaces real tradeoffs in agentic design that don't show up in toy examples.

Agentic Memory

Python Ollama MCP ChromaDB

An MCP server for indexing and storing coding agent memories about repositories. Allows agents to store, recall and utilize past experiences and knowledge across different coding tasks and projects.

04 — Contact

Let's build something together

The best way to reach me is LinkedIn — I actually read it. Always happy to connect around interesting work, ideas, or opportunities.

Built with plain HTML, Tailwind CSS, and AWS S3 + CloudFront. Daily articles powered by an AI agent.