Heet Dalsania

I'm a Computer Science student at Arizona State University's Barrett Honors College, focused on artificial intelligence and machine learning. My work spans biomedical imaging, computer vision, and generative AI.

I currently conduct research at the People and Robots Laboratory (PeRL) under Prof. Pooyan Fazli, and previously worked on label-efficient medical imaging with Prof. Jianming Liang. I'm a Barrett Undergraduate Research Fellow and hold professional ML/AI certifications from Google Cloud, AWS, Oracle, Stanford, and IBM.

Email  /  CV  /  GitHub  /  LinkedIn  /  arXiv

Phoenix, Arizona

• AI-Generated Video Detection
  People and Robots Laboratory (PeRL), ASU — Prof. Pooyan Fazli. Ongoing.
  Developing deep-learning methods to detect synthetically generated video content.
  Focus areas: computer vision and deep learning.
• Label-Efficient Chest X-ray Diagnosis via Partial CLIP Adaptation  pdf / arXiv
  Heet Dalsania. ASU SCENE — Prof. Jianming Liang.
  Zero-shot and few-shot learning for rare disease diagnosis: adapting vision-language models to biomedical imaging with limited labeled data for computer-aided diagnosis.
• Barrett Undergraduate Research Fellowship
  Arizona State University — Prof. Jianming Liang. Completed.
  Pursuing state-of-the-art performance in biomedical imaging and computer-aided diagnosis through advanced deep learning techniques.

• Physiq Python · Computer Vision · ML  code
  AI-powered fitness and body-analysis app that uses computer vision to assess posture and form from a live camera feed, generating personalized workout recommendations in real time.
• Pulse JavaScript · REST APIs · NLP · Sentiment Analysis  live / code
  Stock-news intelligence platform with lexicon-based sentiment analysis (bullish / bearish / neutral) for user watchlists, aggregating Yahoo Finance and Google News and ranking articles by relevance and conviction.
• FADE Python · NLP · LLM · React  code
  Autonomous agent that generates weekly GitHub PR digests and analyzes its own execution trace to classify steps as deterministic, rule-based, or AI-required — cutting costs by up to 99.8% by replacing non-AI steps with plain code.
• NBA MVP Predictor Python · Scikit-learn · Pandas  code
  ML model trained on NBA MVP data and player stats from 1995–2024 to weigh each stat's contribution to MVP shares, achieving 90% accuracy.

• Google Cloud — Professional Machine Learning Engineer
• AWS — Certified AI Practitioner
• Oracle Cloud — Generative AI Professional
• Stanford / DeepLearning.AI — Supervised Machine Learning
• IBM — Applied AI Professional Certificate

• Arizona State University, Barrett Honors College — B.S. Computer Science, 2025–Present. Barrett Undergraduate Research Fellow · GPA 4.0 · Dean's List.
• Skills: Python, JavaScript, PyTorch, TensorFlow, Scikit-learn, Pandas, Flask · Computer Vision, Deep Learning, NLP, CLIP, Few-Shot Learning, Generative AI · AWS, Google Cloud, IBM Watson, Oracle Cloud.
• Activities: Claude Builder Club, Hacker Devils Club, AI Scholars Society.