Hey, I'm Teja Potu
Building intelligent systems at the intersection of Agentic AI, LLMs, and Computational Biology
A bit about me
I'm a PhD student in Computer Science at Florida State University, currently working as a Research Assistant at the National High Magnetic Field Laboratory. My research focuses on building agentic AI systems and applying graph-based machine learning to problems in computational biology and cheminformatics.
I've published in Analytical Chemistry and MDPI Mathematics — and I'm actively seeking PhD internship opportunities in AI/ML and Data Science.
Areas of focus
My work sits at the intersection of agentic AI, scientific computing, and health — building systems that reason, act, and publish.
Health AI & Clinical Systems
Building AI systems that streamline clinical workflows — from multimodal case matching and medical imaging analysis to real-time patient monitoring and referral automation.
Computational Biology
Published research applying graph neural networks and autoencoders to single-cell genomics, and graph-based molecular formula attribution for high-resolution mass spectrometry.
Agentic AI Systems
Designing LLM-orchestrated agents that reason through domain-specific tools — from mass spectrometry pipelines to clinical referral workflows and educational platforms.
Speech & Audio AI
Accent conversion, audio super-resolution, and speech enhancement using autoregressive transformers, GAN architectures, and self-supervised speech representations.
Published research
SCGclust: Single Cell Graph Clustering Using Graph Autoencoders Integrating SNVs and CNAs
T. Potu et al.
MDPI Mathematics, 2025
Internal Calibration without Internal Calibrants in FT-ICR Mass Spectrometry
R.P. Rodgers, C.L. Hendrickson, C.A. Holder Montenegro, A.J. Tello-Rodriguez, T. Potu, et al.
Analytical Chemistry, 2025
Recognition
Agentic Workflow Prize
WinnerMedGemma Impact Challenge
Won the Agentic Workflow Prize for CaseTwin — a clinical decision-support system that matches acute chest X-rays with historical "twins" and uses an agentic workflow to accelerate referrals, turning an hours-long manual retrieval process into a few minutes in rural hospitals.
Education & work
Ph.D. in Computer Science
Jan 2023 – Dec 2027 (Expected)Florida State University
GPA: 3.92
Research Assistant
May 2024 – PresentNational High Magnetic Field Laboratory
Built a graph-based molecular formula attribution engine for high-resolution mass spectrometry. Developing an autonomous LLM-orchestrated system that handles end-to-end spectral data analysis using tool-calling agents.
Graduate Teaching Assistant
Jan 2024 – May 2024Florida State University
Taught Coding Bootcamp and Introduction to Programming (Python).
Researcher — Master's Thesis
Dec 2023 – Dec 2024Single-Cell Clustering
Developed a graph autoencoder framework integrating SNVs and CNAs for single-cell clustering, published in MDPI Mathematics.
Graduate Software Engineer
Aug 2021 – Dec 2022Meeami Technologies
Built speech AI products including noise suppression, echo cancellation, and super-resolution models optimized for edge inference on mobile and embedded devices.
Featured work
Published research, production systems, and open-source projects spanning graph ML, speech AI, deep learning, and AI agents.
CaseTwin
Collapses a 4-hour clinical referral workflow into ~5 minutes by automating historical case matching, specialist facility location, imaging comparison, and referral documentation. Features a medical image RAG pipeline and uses specialized AI models (MedGemma, MedSiglip) for chest X-ray analysis and clinical text processing.
Questory
A quest through knowledge using stories — a narrative-driven educational platform that delivers content through interactive storytelling. Built with a monorepo architecture spanning a TypeScript frontend and Python backend.
PyC2MC: Graph Attribution & Auto-Analysis
Graph-based molecular formula attribution engine for high-resolution mass spectrometry, paired with an LLM-orchestrated analysis layer that interprets spectral data end-to-end. Supports Claude, OpenAI, Gemini, and Ollama as reasoning backends.
SCGclust: Single-Cell Graph Clustering
Lead author on published research integrating SNVs and CNAs using graph autoencoders for robust single-cell clustering. Co-trains graph autoencoder with GCN and GMM for accurate cell subclonality characterization. Consistently outperforms SNV-only and CNA-only methods on simulated and real cancer samples.
More projects
Speech & Audio AI
Data Science & ML
Computer Vision
Reinforcement Learning
Tools I work with
Languages
- Python
- C/C++
- Java
- TypeScript
- JavaScript
ML & Deep Learning
- PyTorch
- TensorFlow
- Hugging Face
- Scikit-learn
- GNN
- GANs
LLMs & Agents
- LangChain
- CrewAI
- RAG
- ChromaDB
- Qdrant
- Prompt Engineering
Infrastructure
- Docker
- AWS
- GCP
- HPC / SLURM
- MLflow
- Git
Let's work together
I'm actively seeking PhD internship opportunities in AI/ML. Open to research collaborations, internship roles, or conversations about agentic AI and computational biology.