Every skill listed here comes from shipped work or a measured project, and each card shows where I used it.
Skills by area
3D vision & robotics
- 3D reconstruction
- Photogrammetry
- Structure from motion
- Multi-view stereo
- COLMAP
- OpenMVG / OpenMVS
- Camera calibration
- SLAM
- ORB-SLAM3
- RTAB-Map
- Sensor fusion (LiDAR, RGB-D, IMU)
- Open3D / PCL
- 3D & 4D Gaussian splatting
- NeRF
Tooliqa, VoxaVerse, PranaTree, brahmGAN, CWRU
Computer vision & ML
- PyTorch
- OpenCV
- YOLOv8
- Mask R-CNN
- OneFormer
- SAM
- DINOv2
- CLIP
- Vision-language models
- GANs
- MLX
Tooliqa, PranaTree, CWRU, KLE Tech
GPU, inference & edge
- Metal
- CUDA
- TensorRT
- ONNX
- NVIDIA Triton
- vLLM
- FP16 / INT8 quantization
- Jetson Orin / TX2
- PyTorch DDP
- Pipeline parallelism
- WebGPU
VoxaVerse, IdeaBoxAI, PranaTree, Tooliqa, brahmGAN
NVIDIA AI stack
- NeMo
- VSS (Video Search & Summarization)
- Omniverse
- Triton
- TensorRT
- Jetson
IdeaBoxAI, PranaTree
Vector search & data
- Vector databases
- FAISS / HNSW
- Milvus
- Qdrant
- VectorDBBench
- gRPC / Protocol Buffers
- FastAPI
- PostgreSQL
- Redis / Celery
IdeaBoxAI
Agents & LLMs
- Model Context Protocol (MCP)
- LangGraph
- LangChain
- RAG / GraphRAG
- Tool-routing evaluation
- llama.cpp
- OpenAI, Anthropic and Gemini APIs
IdeaBoxAI, PranaTree
Security
- OAuth 2.1 / JWT / OIDC
- TLS / mTLS
- SSRF defence
- Tamarin / ProVerif
IdeaBoxAI, Axhara
Languages & delivery
- Python
- C++
- Swift
- Rust
- TypeScript
- C#
- Metal Shading Language
- Docker
- GitHub Actions
- Xcode Cloud
- pytest
- ruff
- mypy
All of the above