
Alex NovaraSoftware engineer building reliable web platforms and practical AI tools
Alex is a fictional software engineer from the demo seed. Ten years of web platforms for logistics, retail and publishing, and lately agents, retrieval and evaluation.
- Would Alex Novara fit an AI engineering role?
- What could Alex Novara automate for my company?
What I do
Software systems
Web applications and their backends, built to be maintained long after launch.
Integrations & automation
Connecting the tools a team already uses and taking repetitive steps off its hands.
AI engineering
My current focus: useful AI features in real software, explored in independent R&D.
This website is open source and built with an agentic engineering workflow.Source code on GitHub
Selected work
Reading assistant
March 1st, 2024 – present
- Problem
Readers searched an archive of twenty years by keywords and rarely found the article they remembered.
- My role
Technical lead
- Result
Four in five test questions are answered with a correct source; the suite runs on every change.
- Python
- LangGraph
- pgvector
- PostgreSQL
Editorial workbench
June 1st, 2022 – September 30th, 2023
- Problem
Editors juggled three tools and a shared spreadsheet to publish one story.
- My role
Lead engineer
- Result
Publishing a story takes one tool and half the steps.
- TypeScript
- Next.js
- PostgreSQL
Store replatforming
April 1st, 2019 – February 28th, 2021
- Problem
Every change to the old store risked the checkout.
- Result
Releases went from monthly to daily without a checkout outage.
- TypeScript
- Node.js
- React
- GraphQL
- PostgreSQL
Experience
11 years of professional software development. Currently: Staff engineer, Lumen Press (full-time job); Independent consultant, Quill Studio (own business); Maintainer, Open Shelf Collective (side project).
- Core
- TypeScript · Node.js · React · Next.js · PostgreSQL · PHP
- AI engineering focus
- Python · LangGraph · pgvector · Model Context Protocol
PDF updated
Now
Working on
Taking the reading assistant to every reader
Moving the fictional reading assistant from a test group to all readers, with the evaluation suite as the release gate.
- Retrieval
- Evaluation
Permissions for tool-using agents
Turning the designed tool permission model into a small library.
- Agents
- Tool use
- Guardrails
Statistics for evaluations
Learning how many evaluation questions are enough to trust a difference between two versions.
- Evaluation
Exploring
- Foundations
In progress4/8 resources completed
- Embeddings1/3
- Retrieval systems
In progress1/10 resources completed
Services
AI assistant prototype
A working assistant over your own documents in four weeks, with an evaluation set that shows how well it answers.
Architecture review
A two-day review of a web platform: what holds, what will break first and what to change.
Evaluation setup
An evaluation suite for an existing AI feature, so changes to prompts and models are measured, not guessed.
Have a project or a role in mind?
Write to alex@novara.example.test