AI Engineer & Researcher
Building models, startups, and breakthroughs in machine learning

📍 Tucson, Arizona | Portfolio | GitHub

AI Engineer and Researcher with 3+ years of experience in applied machine learning and deep learning. My work spans healthcare AI systems, multi-agent architectures, multimodal LLMs, and protein design. I’ve co-founded startups, published preprints, and filed patents at the intersection of AI, health, and science. Passionate about open-source, research collaboration, and building impactful AI solutions.

• University of Arizona – M.S. in Information Science: Machine Learning [Distinguished Scholor]
• National Institute of Technology – B.Tech in Electronics and Communication Engineering

Jan 2025 – Present | Tucson, Arizona

• Collaborating with a leading semiconductor company on Generative AI, RAG, multi-agent systems, and AI-driven simulations.
• Built digital phenotyping pipelines using mobile-sensing platforms and deployed RADAR-Base on AWS for real-time data ingestion.
• Engineered a multi-agent architecture unifying wearable data, smartphone signals, clinical text, and surveys into a real-time analytics framework.

• Co-founded an AI-driven startup reimagining protein and antibody design for drug discovery.
• Led deep learning pipelines with academic labs, achieving 6.53% improvement on charged proteins and 4.14% overall gain in sequence recovery.

• Designed a novel multi-agent architecture for short-form video understanding, combining OpenAI agents, OpenCV, and Transformers.
• Built a custom model trained on 5,058 influencer videos, achieving SOTA results.

• Developed an automated portfolio analysis system, reducing consultant effort from 40–50 hours to minutes.
• Built an AI workbench with multi-agent systems, improving efficiency by 70% and boosting accuracy from 6% to 72%.

• Built a clinical AI system with ontology fusion (UMLS, SNOMED CT, ICD-10) and RAG for evidence-grounded diagnosis.
• Achieved 0.7989 Macro F1 (+11.1% vs. BioClinicalBERT) and 0.9155 AUROC on 8,604 MIMIC-IV notes.
• Awarded Best Graduate Project and ReaP Research Grant.

• AI system to detect comets in NASA SOHO/LASCO images using EfficientNet-B0 and difference imaging.
• Achieved 97.7% accuracy, 98% precision, 99% recall.
• Deployed as a real-time Gradio app with 2s inference per sequence.

• Developed an AI-powered platform with real-time environmental analysis, RAG-based insights, and interactive 2D/3D simulations.

Languages: Python, C++, C, SQL, JavaScript
ML/DL: Transformers, GNNs, RAG, Fine-tuning, Multi-Agent Systems, XAI
Frameworks: PyTorch, TensorFlow, OpenCV, FastAPI, Streamlit, Hugging Face, MATLAB
Tools: Docker, AWS, GCP, Kubernetes, Firebase, MongoDB

• Forthcoming arXiv preprint (Q4 2025) on novel multimodal LLM fine-tuning architecture (Quelea).
• Forthcoming arXiv preprint (Q1 2026) on protein design with fusion architectures (IonTheFold).
• Patent holder for an autonomous self-landing rocket control system (Pub. No. 47/2025).