Skills
Platforms, libraries, and frameworks used across deep learning, applied statistics, generative AI, and MLOps.
Deep Learning & Core Modeling
Libraries & Frameworks
- PyTorch / PyTorch Lightning — dynamic computation graphs, multi-GPU training
- JAX / Flax / Equinox — high-performance numerical computing & auto-vectorization
- TensorFlow / Keras — legacy enterprise deployments & production serving
Platforms & Infrastructure
- Hugging Face Hub & Transformers — foundation model ecosystem
- DeepSpeed & Megatron-LM — distributed large-scale training optimization
Applied Mathematics & Statistics
Libraries & Frameworks
- SciPy & NumPy — advanced scientific computing & matrix operations
- Statsmodels — parametric/non-parametric statistical modeling & hypothesis testing
- SymPy — symbolic mathematics & algebraic derivations
Platforms & Infrastructure
- Stan / PyMC — Bayesian inference & probabilistic programming
- NetworkX — graph theory & complex network analysis
Generative AI & LLM Engineering
Libraries & Frameworks
- vLLM & TensorRT-LLM — high-throughput, low-latency LLM inference serving
- LangChain & LlamaIndex — orchestration for agentic workflows & RAG pipelines
- PEFT & bitsandbytes — parameter-efficient tuning & model quantization
Platforms & Infrastructure
- Ollama / Triton Inference Server — production model serving & hardware acceleration
- Ray / Anyscale — distributed computing for scalable AI workloads
MLOps, Data & Infrastructure
Libraries & Frameworks
- MLflow & Weights & Biases (W&B) — experiment tracking, model registry, & lineage
- Feast / Hopsworks — enterprise feature stores for real-time inference
- Ray Tune & Optuna — distributed hyperparameter optimization
Platforms & Infrastructure
- Kubernetes & Docker — containerization & orchestration
- Apache Spark / Databricks — large-scale data processing & lakehouse architecture
- Airflow / Kubeflow — ML pipeline orchestration & CI/CD workflows