Python: Primary language for deep learning research, FastAPI services, scientific simulation, and data pipelines.
C++: High-performance systems programming, custom Vulkan ray tracing engines, and latency-critical compute.
TypeScript: Strictly typed full-stack interfaces, Next.js, and browser-native simulations.
Cython: C-extensions and acceleration for inner simulation loops and numerical operations.
PyTorch: Training computer vision architectures, custom neural predictors, and embeddings.
ONNX & TensorRT: Quantized edge model export and sub-millisecond inference acceleration on RTX-class GPUs.
SciPy & FilterPy: Extended Kalman Filtering, sensor fusion, numerical integration (RK4), and control theory (LQR).
FastAPI & Node.js: Asynchronous REST APIs, streaming responses, and high-speed ML model inference runners.
PostgreSQL & PostGIS: Spatial indexing, GeoAlchemy2 geospatial queries, and composite index optimization.
Docker & Docker Compose: Multi-container microservices, reproducible ML environments, and CI deployment stacks.
Vulkan: Low-overhead GPU pipelines, Ray Tracing KHR extensions, and indirect command dispatch.
CUDA: Massively parallel GPU computing, matrix arithmetic kernels, and custom ray allocation passes.
Mitsuba 3: Dual-pass renderer, G-buffer packing, and synthetic dataset generation for path tracing models.