# Ford Lascari

## Education

### Georgia Institute of Technology

Master of Science in Computer Science

August 2021 - December 2022

Specialization in Computer Graphics

- Coursework: Deep Learning for Text, Game AI, Blockchains, Computer Animation, Shape Grammars.

### Georgia Institute of Technology

Bachelor of Science in Computer Science

August 2016 - May 2021

Specializations in Modeling/Simulation and Computer Intelligence

- Coursework: Deep Learning, Machine Learning, Computer Simulation, Numerical Analysis, Robotics.

## Experience

### Lascari AI

Independent Applied AI Consultant

August 2024 - Present

- Help businesses implement and improve large language model and agent systems.
- Lead agent development for voice-based logistics and adaptive customer research.

### Width AI

AI Engineer / Full Stack Engineer · Contract

July 2023 - August 2024

- Built full stack applications powered by large language models.
- Developed TorqAgent, a custom no-code agent-building framework for Intellibridge.

### VRSEN AI

Technical Project Manager / Lead AI Developer · Contract

March 2024 - July 2024

- Led development teams building multi-agent systems across a range of use cases.
- Handled project scoping, technical delivery, sales, customer success, hiring, and training.

### Morningside AI

AI Engineer / Full Stack Engineer · Contract

October 2023 - January 2024

- Built generative AI applications using the OpenAI Assistants API.
- Delivered the initial MVPs for ReadyRNs, an AI-assisted nursing education platform, and Agentive, a low-code agent builder.

### Georgia Tech Research Institute

Machine Learning Research Assistant

June 2020 - December 2022

Electro Optical Systems Lab · Electronic Warfare Division

- Developed machine learning models and algorithms, including neural networks and genetic algorithms.
- Worked on aerial object detection, time series classification, and classification using particle filters.
- Analyzed simulation outputs and created visualizations and research presentations.

## Projects

### Bubba AI

Lascari AI · Hey Bubba

Voice agents for owner-operator truck drivers.

- Led agent development and built outbound agents that called carriers to negotiate freight rates.
- Designed orchestration infrastructure for automated load booking, real-time load search, and internal operations.

Voice agents · Logistics · Agent orchestration

### Dynamic Interview Agent

Lascari AI · Deep Need

Adaptive interviews for customer research.

- Built a conversational agent that adjusted interviews to participant responses, with support for hundreds of concurrent sessions.
- Used clustering to identify patterns in customer responses.

Conversational AI · Customer research · Clustering

### House Hunter

Personal project

A real estate search tool built around better backyards.

- Scraped Trulia listings and images, then used vision models to identify backyard photos.
- Trained a ranking model on personal preferences to filter properties by outdoor space.

LLM vision · Web scraping · Ranking models

### X Replier

Personal project

A tool for finding relevant conversations on Twitter/X.

- Collected tweets from curated lists and trained an engagement classifier on personal interaction history.
- Added image and video analysis for context, with Celery workers for asynchronous processing.

Classification · Media analysis · Celery

### ReadyRNs

Morningside AI

AI-assisted learning for nurses.

- Built the initial MVP with specialized assistants using the OpenAI Assistants API and support for uploaded learning materials.

OpenAI Assistants API · Education

### Agentive

Morningside AI

A low-code platform for building autonomous agents.

- Built the initial MVP on the OpenAI Assistants API, designed to support thousands of users.

Agent builders · OpenAI Assistants API

### TorqAgent

Width AI · Intellibridge

A custom framework for building agents without code.

- Provided full stack development and connected models and APIs to tools agents could use to complete tasks.

Full stack development · Agent tools

### Synthetic Polling

Personal project

Simulated polling with language models.

- Used Montana census and salary data to build a stratified dataset of personas, queried them with OpenAI GPT models, and compared the results with real polls.

Synthetic data · OpenAI GPT · Polling

### Tales For Tots

Web application

AI-generated children's picture books.

- Used GPT-4 for stories and image prompts, Stability and Replicate for image generation and manipulation, and Turn.js to display the books.

GPT-4 · Image generation · Turn.js

## Skills & Interests

### AI & Data

Python, TensorFlow, PyTorch, Keras, Pandas, model fine-tuning, SQL

### Application Development

Flask, FastAPI, JavaScript, HTML, Java, C#, C/C++, Processing, p5.js

### Technical Leadership

Project scoping, team leadership, hiring, training, research, presentations, public speaking

### Beyond Software

Welding, machining, workshop tools; tenor and alto saxophone, beatboxing; golf, training, running