Skip to content
View tyhobbs's full-sized avatar
1 follower · 1 following

Block or report tyhobbs

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
tyhobbs/README.md

Hi, I'm Tyler

Machine learning engineer and applied scientist, finishing an M.S. in Data Science at the University of Virginia in December 2026. Before data science I spent four years in proteomics and genomics labs.

I build models and then check whether their numbers mean what they appear to. My projects tend to end with an audit: a simple baseline that explains a headline metric, a domain gap a public benchmark hid, or a result that didn't survive more random seeds.

Currently: Learning Javascript, React, and Next.js.

Portfolio · Case studies · Hugging Face · LinkedIn

What I work with

Languages

Python SQL R Bash

Deep learning

PyTorch TensorFlow Hugging Face LoRA / PEFT U-Net++ PPO

Data and evaluation

pandas NumPy scikit-learn Plotly Multi-seed analysis LLM-as-judge

Shipping and infrastructure

Flask Gradio Git Linux HPC (SLURM) Vercel

Featured work

LoadBrief — a LoRA fine-tune of Llama 3 8B that writes structured athlete load-management briefs, and a paper on what its accuracy actually measured. Case study

Automated glaucoma screening — optic disc and cup segmentation from retinal images, and what happened when a public-data model met real clinic images. Most Innovative Analytical Solution award. Case study · Live demo

Deep RL stock trading — a PPO trading system with news sentiment, stress-tested across random seeds and stock universes. Case study

Pinned Loading

  1. finrl-dashboard finrl-dashboard Public

    Dashboard for FinRL Deep Reinforcement Learning Project

    HTML 1

  2. FinRL_Deep_Reinforcement_Learning FinRL_Deep_Reinforcement_Learning Public

    A progressive DRL stock trading system built on FinRL, benchmarking four model generations across VGG CNN and Transformer architectures on three capital levels. Trained on up to 50 NASDAQ tickers (…

    Jupyter Notebook

  3. LoadBrief LoadBrief Public

    Fine-tuning Llama 3 8B to convert athlete-monitoring data into structured load-management briefs — grounded in published sports-science rules.

    Python

  4. Rjashby1/Automated-Glaucoma-Screening-Using-AI-Enhanced-Ophthalmoscopy Rjashby1/Automated-Glaucoma-Screening-Using-AI-Enhanced-Ophthalmoscopy Public

    Reproducible optic disc/cup segmentation and public-to-clinical transfer study for AI-enhanced ophthalmoscopy.

    HTML 2