John Patrick Capocyan
About
Hi! I'm currently a Computer Science student at The University of Texas at Austin. I am deeply passionate about Deep Learning and even more dedicated to advancing the fields of Computer Vision and Trustworthy AI. I believe that applying these concepts in critical fields like healthcare is one of the most challenging barriers in the AI world. So, my primary focus of research is algorithmic design and optimization of fine-tuning models.
I had a constantly-evolving path through computer science in that I started learning HTML/CSS in 8th grade and took on an interest in AI and Machine Learning during the summer between my sophomore and junior year of high school. I have a technical background in web development and AI engineering/research -- largely because I love learning new things about math, statistics, and neural networks!
Education
B.S. in Computer Science
The University of Texas at Austin, GPA: 3.9/4.0
I'm currently an Officer Board Member of Texas Luminescence, an AI engineering organization. Last semester, I
was a project team lead, directing 11 members across frontend, backend, and ML engineering subteams to
build HealthLens, a Computer Vision + NLP powered app to automate dermatology screening. I'm also a member of the Directed Reading Program (DiRP) and
Machine Learning & Data Science Club, and I also play for UT Chess Club's A Team.
Oh yeah, I also run a coffee business on West Campus for UT students. I post
under the account of @jpcappuccino on Instagram and TikTok! :)
High School Diploma
Clements High School, Sugar Land, Texas, GPA: 4.3/4.0
I graduated Summa Cum Laude, being in the top 10% of my class while developing my passion for AI and research. I was also a Co-Captain of my school's Varsity I Golf Team, member of CS National Honor Society, and a coach/competitor for our school's eSports Chess Team!
Experience
Undergraduate AI Research Assistant
2026: Cenik Lab in Computational Biology
- Trained convolutional and transformer-based architectures to predict RNA efficiency from ribosome profiling.
- Scaled DNA context window from 13k to 500k base pairs (38x increase), improving biological signal capture.
- Utilized Linux and SSH to access TACC’s A100 GPUs, accelerating large-scale training for deep learning models.
- Advised by Dr. Can Cenik and Master's student Logan Persyn
Undergraduate Research Assistant
2025: Baker Group in Quantum Computing
- Processsed 10k+ quantum circuit samples using Pandas and NumPy, identifying 50+ critical circuit encodings.
- Generated 100+ Matplotlib visualizations to analyze relationships between circuit fidelity and fault tolerance.
- Advised by Dr. Johnathan Baker and PhD students Cordell Mazzetti and Sayam Sethi
Founder
2025: Validly
- Developed an online platform for startups and business ideas to automate their market research
- Lead 3 developers and 1 growth intern, overseeing all technical development and advertisement, leading to 10,000+ online impressions, 100+ social media followers, and 10 early-access users from 44 waitlist sign ups
Software Engineering Intern
2024: iAnswer
- Implemented RAG to fine-tune a LLM on 1000+ anonymized clinical PDFs, improving text querying by 20%.
- Built a PyTorch-based embedding pipeline producing dense vectors (768-dim) optimized for semantic retrieval.
- Connected React components with Node.js backend through REST APIs, improving response efficiency by 50%.
Accomplishments
In my life, I've done a few things of which I'm particularly proud, largely because they required several years of sustained effort:
- In 2025 (fall), I received a total of $1,500 in research funding from the UT Austin department to support my independent research in LLM interaction
- In 2025 (summer), I completed the 100,000 steps in 24 hours challenge, hiking 47.42 miles and burning 9080 calories in 23 hours and 54 minutes
- In 2025 (spring), I was admitted to The University of Texas at Austin's B.S program in computer science and UCLA's B.S. program in computer science
- In 2024, I finished my second independent AI research project and presented it at the 2024 IEEE MIT URTC in front of other chair/faculty members
- In 2023, I finished my first independent AI research project and presented it at the 2024 IEEE ISEC at Princeton University in front of other chair/faculty members
- In 2018, I was represented Texas in the Barber Chess Tournament's Middle School Division (1/50 competitors for each state)
- From 2015 to 2018, I was State Champion for my grade level in the K-12 Texas State Chess Tournament
- In 2015, I played for the U.S. Junior National Team in the World Youth Chess Championship in Greece
Selected Awards
- 2026: 8th Place in the 6th Annual Energy AI Hackathon, UT Austin PGE
- Developed a gradient-boosted decision tree model predicting 3-year cumulative oil output with a 0.106 error metric.
- Ranked among 40+ teams and 200+ undergraduate and graduate participants.
- 2025: Undergraduate Research Fellowship ($1000), UT Austin
- 2025: Faculty Research Grants Scholarship ($500), UT Austin
- 2025: Associate Membership, Sigma Xi Scientific Research Honor Society
- 2025: MIT URTC Paper Officially Published as a Conference Proceeding in IEEE
- Paper titled "Enhancement Framework for Vision Transformers in Data Limited Cancer Detection"
- Was 1 of 58 sole authors at the conference (250+ papers).
- 2024: IEEE ISEC Paper Officially Published as a Conference Proceeding in IEEE
- Paper titled "Implementation of Convolutional Neural Networks for Classifying Lung Cancer Types from Histopathological Images"
- Was 1 of 34 sole authors at the conference (150+ papers).
Hobbies
- Golf
- Chess
- Photography
- Coffee Making
- Travel
- Reading AI Research Papers
Note: I publish under "J.P. Capocyan," though my legal name is "John Patrick Capocyan."
Contact
Austin, TX
Personal Email: johncapocyan@gmail.com
Research Email: jcc6945@eid.utexas.edu
LinkedIn: linkedin.com/in/johnpatrickcapocyanai
GitHub: github.com/J7P16
Instagram: @j7p16