MS Data Science · University of Arizona
Tanishk Singh
Data Scientist & ML Engineer — efficient systems, from CUDA kernels to production RAG pipelines.
Focused on ML systems efficiency, federated learning, and hardware-aware inference — turning research into things that actually run.
156x
CUDA kernel speedup on Mamba SSM inference (A100)
70%
less communication overhead in federated fine-tuning
54,000
students reached through WWF's Wild Wisdom Challenge
Featured Projects
Research and applied work across federated learning, CUDA/systems optimization, and production LLM pipelines.
Toolkit
What I reach for when building and researching.
Languages
- Python
- C++/C
- Java
- R
- SQL
ML/DL Frameworks
- PyTorch
- TensorFlow
- CUDA
- scikit-learn
- Pandas
- NumPy
ML/AI Techniques
- Deep Learning
- Federated Learning
- NLP
- LLM
- RAG
- LoRA/PEFT
Backend & Infra
- SpringBoot
- Docker
- Kafka
- PySpark
- MongoDB
- Postgres
- pgvector
- HPC
- MLOps
- A/B Testing