Aryan Singh
GPA: 4.0
- Minors: Entrepreneurship and Management, Applied Math and Statistics
- Coursework: Machine Learning, Deep Learning, Algorithms, Probability Theory, Linear Algebra, Optimization, Systems Programming, Quantum Computing
- Cofounded and built a production-grade multimodal ML platform for neurological screening using speech and language models, trained on 3,000+ hours of clinical and public-health audio (12TB+).
- Designed and deployed end-to-end ML infrastructure spanning data ingestion, feature extraction, model training, evaluation, and batch inference, delivering 25–30% accuracy gains over baseline cognitive screening tools.
- Deployed pilots with state health agencies across 8 US states; enabled scalable deployment for 50,000+ screenings/year.
- Raised $150K+ seed funding by communicating model performance, uncertainty, and deployment constraints to clinicians and policymakers.
- Built ML-driven data pipelines for analyzing conductive polyaniline nanofibers in nerve and tissue regeneration research.
- Processed and modeled 120,000+ high-resolution experimental micrographs using custom computer vision workflows, reducing material optimization cycles by 80%+ and compressing research iteration time from months to days.
- Developed genome-scale ML systems for CRISPR off-target prediction using experimental Geminin-tagged Cas9 (gCas9), enabling ultra-high-specificity gene editing for neurodegenerative disease research.
- Applied rigorous statistical modeling to 10+ TB of sequencing data, cutting in-silico screening time from weeks to <6 hours.
- Work awarded Top Prize at Rice Neurotransmitter Research Competition Nationals.
Built a full-stack, modular quantum circuit simulator with multiple execution backends (statevector, density-matrix, stabilizer, tensor-network), designed for correctness, extensibility, and high-performance experimentation.
- Integrated AI/ML-driven analysis to model noise, detect failure modes, and predict breakdowns in quantum circuits.
- Dynamic visualizations for diagnosing behavior under entanglement growth and hardware constraints.
Built a scalable ML evaluation framework for embodied AI agent reasoning, running 2,400+ large-scale simulations to stress-test model failure modes, decisions, and false-beliefs using Python workflows (NumPy, TensorFlow, pandas, joblib).
Top 0.5% Nationally
Top 50 Nationally
- Advised the Mayor's Office, Houston City Council, and Texas legislative staff on youth mental health and STEM access initiatives impacting multiple congressional districts.
- Hosted 2-season podcast (100K+ views online) featuring the mayor, city officials, and educators.
- Contributed to proposals influencing $7M+ in federal funding for youth mental health & STEM mentorship programs.
Languages
Python (PyTorch, TensorFlow, NumPy, Pandas), C++, MATLAB, JavaScript, Java, SQL, Bash
Full-Stack
React, Angular, HTML, CSS, REST APIs, microservices, server-side development, API design
ML & Data
Deep learning, transformers and attention, multimodal models, computer vision, NLP, speech and audio modeling, model evaluation, Hugging Face, OpenCV
Systems
Linux/Unix, Git, Docker, distributed data pipelines, parallel computing, relational databases, large-scale data processing