About Me
I'm a technologist working at the intersection of accessibility, AI, and language systems. My work centers on understanding how technology can be designed, implemented, and improved to create more equitable access to information.
What I Do
At Rochester Institute of Technology, I coordinate real-time captioning services while conducting applied research in natural language processing and speech recognition. I've built production systems ranging from equipment management applications to experimental transcription enhancement tools.
My technical work spans multiple domains: I develop systems in Python, fine-tune language models on specialized corpora, and implement speech recognition pipelines. But the through-line in all of it is a focus on understanding what people actually need from technology—and then building solutions that deliver on that promise.
I'm particularly interested in the gap between what commercial technology claims to do and what it actually delivers in practice. I've spent considerable time reverse-engineering systems to understand their true capabilities, conducting accuracy analyses on speech recognition platforms, and advocating for evidence-based adoption of accessibility technologies.
This Site
This site serves as a repository for my technical work, research findings, and analysis of accessibility technology implementations. You'll find documentation of systems I've built, corpus analysis results, and writing on the state of real-time transcription and captioning technology.
I believe technology should be evaluated rigorously, implemented thoughtfully, and continuously improved based on actual performance rather than vendor promises.
Technical Focus
Machine Learning & NLP
Fine-tuning large language models (Llama, Qwen) with LoRA adapters, building custom NLP pipelines for transcript analysis, and developing domain-specific language models.
Systems Programming
Building systems with focus on performance and reliability. Real-time processing pipelines for transcription and speech recognition workflows.
Speech Recognition
Implementing and optimizing Vosk, Kaldi, and Azure STT systems for academic transcription. Building hybrid ASR pipelines that combine automatic speech recognition with intelligent text prediction and correction.
Web Applications
Full-stack development with Python (Flask, FastAPI), React, and modern web technologies. Building production systems with proper security, authentication, API integration, and real-time updates via WebSockets.
Key Projects
- Kiosk: Built for $9, it replaced paper-based tracking for tens of thousands of dollars in departmental equipment. A cross-departmental effort, still running in production.
- Event Portal: A prototype to replace Excel-based event calendars, with a virtual pet for engagement. Taught me requirements gathering through iteration.
- TAP: Hand-coded TF-IDF alongside a professor to make years of transcript data searchable. The project that got me into NLP.
Experience
Real-Time Captioning Coordinator III
Rochester Institute of Technology
Leading implementation of AI/ML solutions for accessibility services. Conducting funded research on NLP analysis of academic transcripts. Developed multiple production systems serving campus-wide operations.
Captioning Team Specialist II
Rochester Institute of Technology
Developed comprehensive training programs and technical documentation. Began research into ML applications for transcript analysis.
Real-Time Captionist I
Rochester Institute of Technology
Mastered TypeWell's advanced Extreme mode through reverse-engineering. Achieved 200+ WPM transcription speed with high accuracy.
Technical Skills
Languages
Python, JavaScript/TypeScript, MATLAB, SQL
ML/AI Frameworks
PyTorch, Transformers, PEFT, Hugging Face, NLTK, sklearn
Speech Recognition
Vosk, Kaldi, KenLM, Azure STT, Deepgram
Web Development
FastAPI, Flask, React, WebSockets, Apache
Systems & Infrastructure
Apache, Performance Optimization, API Integration
Tools & Platforms
Git, Docker, WSL/Ubuntu, Cloud GPU (Vast.ai), Linux
Education
Bachelor of Arts in English
Binghamton University, Binghamton, NY
Let's Connect
I'm always interested in discussing technology, accessibility, machine learning, and systems programming. Feel free to reach out!