Embedded Deep Learning Engineer.
With ~6 years of industry experience below are major areas I've worked on:
Deep Learning:
- Various architectures: VGG, MV1, MV2, ResNet, SSD, SqeezeDet, SqeezeNet
- Model designing, training, optimisation, quantization, pruning
- Framework: TesnsorFlow (including tflite, tfmicro), Caffe, ONNX, Darknet
- NN compiler development
- Architecting docker container based ML training framework
- ML HW platforms: Google coral, Intel movidius, Rpi, AWS DeepLens, Xilinx and Lattice FPGAs, etc
- ML Cloud Platforms: GCP (AutoML, Vision API), AWS (Sagemaker, Rekognition)
Firmware Development:
- Platform: Tensilica based Analog Audio Processor chip
- IoT based gateway and edge devices
Networking:
- Kernel's mac driver customisation for mesh networks
- l2/l3/Application layer filtering
- Mobility solution development
- OpenWRT router OS customisation
- Captive portals & radius servers