Hi, I'm Jordan


Jordan Romano



Machine Learning

Machine Learning



Achieved top 1% ranking on Hugging Face for the LunarLanderv2 competition; a physics-based simulation project where an AI agent is trained to autonomously land on the moon.

Scored in the top 2% on Kaggle for the Titanic ML competition. The goal of this competition is to train a machine learning model that can accurately predict which passengers survived the shipwreck.

Transit Information Assistant

Transit Information Assistant



AskLynchburg - a bilingual information assistant for bus riders, developed independently and later acquired by Ineo Systrans USA (EQUANS).

Providing instant local insights, it was initially designed for the Lynchburg, VA community, with geo-location features that work globally.

- Nearby bus stops with real-time GTFS arrival times, fare prices, and bus policies.
- Nearby parks and open restaurants, including Yelp ratings
- Speech-to-text, Natural Language Processing

Customer Service Solution

Customer Service Solution



AskGLTC, a bilingual customer service chatbot & WordPress plugin leveraging GTFS data and ML frameworks.

- Assisted 300+ riders through integration with the Greater Lynchburg Transit Company website (Jan. 2023 – Aug. 2023).

- Twilio SMS integration, bus arrival times at nearby stops, speech transcription, GTFS feed integration, pdf forms, videos, tutorials and AI chat functionality.

Presented to Board of Director members such as the Lynchburg Deputy City Manager, Senior VP of Liberty University, Regional VP of TransDev and GLTC management.

Recognized by the DRPT for improving rider experiences. (p. 33)

Awarded the William Millar scholarship by the American Public Transportation Foundation (APTA) in connection with this project.

Web3 Development on Solana

Web3 Development on Solana



AutoSol, a 1000 unit Solana NFT collection that sold out in 30 minutes to 700 investors and generated $26,000 in profit.

Led the project and took charge of marketing in collaboration with a developer and an artist.

This collection currently trades on major SOL marketplaces such as SolSea, TensorTrade and Magic Eden.



Skills


  • Python

    Experience with Python in application development and machine learning. Used in the AskLynchburg and AskGLTC projects.


  • Flask

    Experience with the Flask micro web Python framework for the deployment of web applications. Used in the AskLynchburg and AskGLTC projects.


  • PyTorch

    A Python machine learning framework for finetuning and deploying ML models. Used in the AskLynchburg and AskGLTC projects.



  • JavaScript

    Developing advanced web functionality for the AskLynchburg and AskGLTC projects. React.js for AutoSol's deployment and the development of this portfolio site.


  • WordPress

    Maintaining and deploying WordPress sites, some experience with plugins: Built a simple WordPress Plugin for the AskGLTC chatbot.


  • REST APIs

    Application Programming Interfaces (APIs) such as Transit APIs, Google Places, Twilio and others.



  • AWS Cloud

    Hosting and deploying using Amazon Web Services: EC2, VPC, LightSail, Route53 and CloudFront.


  • OpenAI Gym

    An API standard for reinforcement learning environments. Used for 5 HuggingFace projects.



About

I strive to maintain a focus on machine learning, cloud hardware and systems integration. Previously, I've worked with NAVINEO, a CAD/AVL transit solution, where I handled its integration into cloud environments and physical hardware like MDTs, VLUs, and tablets. Currently I assist clients with solutions for cloud hardware configurations that are scalable and ideal for HPC, LLMs and Deep Learning.

Independently, I've developed tools like AskLynchburg, an assistant for bus riders, and AskGLTC, a customer service chatbot that helped hundreds of riders. These projects involved Python, Flask, and machine learning frameworks like PyTorch, allowing me to build real-world solutions with a direct impact.