Freelance · Software & AI
Time Tracker, HireSight and Mora. New products explored with the Edgeworks team.
My collaboration with Edgeworks began through Salvatore Iannola, a senior software engineer and one of the company’s administrators. I wanted to put my knowledge into practice, while the team wanted to explore new products and solutions to show clients. Those two needs brought us together.
My role was to experiment with those solutions and help build them. I worked with Salvatore across Time Tracker, HireSight and Mora: he provided direction, feedback and corrections as the projects took shape. It was an interesting collaboration that also let me get to know the wider Edgeworks team.
Time Tracker is the most important result for me and is still in use as of October 2026. HireSight has also been used internally by Edgeworks and shown to interested clients, while Mora was an early prototype for classifying incoming emails and preparing draft replies. Each project was a different way to turn an everyday need into software.
Salvatore encouraged me to keep improving and gave me constructive criticism alongside positive feedback. That ongoing exchange mattered more than any single correction: I could try ideas, discuss them with an experienced engineer and refine the work.
01 / The project
Time Tracker grew out of an earlier experience at Marsilea, where I had built a time-tracking tool with Google Sheets and Google Apps Script. Working with that setup had been frustrating. When Edgeworks also needed an efficient way to record hours, we decided to build a new application from scratch.
02 / The project
We wanted a simple, intuitive interface that worked well on both phones and desktops, backed by a database. Next.js and React provided the interface, with TypeScript, Supabase and Recharts supporting the application, data and charts. The tool brings together time entries, projects, clients and reporting, with authentication and Excel exports.
03 / The project
The main improvement for me was having a workflow tailored to the task. In the spreadsheet, entries could become disorganised and some errors were difficult to handle. Time Tracker gave us more control over how information was entered and managed in the database, through dedicated operations and an interface designed around recording hours.
It is a straightforward tool, and that is what I like about it: it does the job and remains in use. Building it was satisfying because an experience with a cumbersome workflow became the starting point for something more practical.
04 / The project
HireSight began with the need to receive CVs, analyse them and organise them so HR could have an overview of both candidates and employees.
05 / The project
React, TypeScript, Python and Supabase connect the interface, data and AI-assisted analysis.
06 / The project
A recruitment platform developed in my work with Edgeworks. It helps organise job positions and candidates, with AI-assisted CV analysis.
07 / The project
The idea was to bring CVs and skills into a shared database, useful for preparing interviews and understanding which capabilities were available for different projects.
08 / The project
The interface connects candidate details, match scores and the reasoning behind them, so recruiters can review the information in context.
09 / The project
Mora was a very early MVP: an email orchestrator that prepared automatic draft replies based on specific company requests. My experience with it was an initial experiment, with a small RAG system to retrieve relevant context for the drafts.
The prototype explored repetitive, standard communications: information requests, requests for estimates and other categories of incoming mail. It also explored distinguishing spam from other messages and switching language models according to cost and other requirements. The aim was to test a product that could save companies time when handling recurring enquiries.
The interesting part was connecting an incoming request to information that could help answer it. Retrieval provided context to the language model; the result was a draft to review. The scope of this work was to test that idea in an early prototype.
Edgeworks now describes Mora as a Python-based email workflow that retrieves unread messages, classifies them and generates contextual drafts, with Gmail API integration. Its product page provides a reference for the workflow; my contribution here concerns the early MVP.
For the curious
The tools mentioned in the story, with documentation, guides and source code to help you build something of your own.
If you have a question about this project or an idea to discuss, write to me. I’m happy to share what I’ve learned.
Ask me a question