Experienced embedded software engineer working on Embedded Systems and Deep Learning to enable vision and voice-based machine learning algorithms on low-power FPGA and edge embedded devices. ~8 years of experience consists in writing, debugging, and optimizing software/firmware for embedded devices.
+91 9409 24 93 94
kishanpgondaliya@gmail.com Ahmedabad, Gujarat, India
Languages:
Frameworks:
Dev Tools:
HW Platform:
Cloud (GCP):
Cloud (AWS):
Other:
C, Python, C++
Tensorflow (TFlite, TFmicro), Keras, Caffe, Darknet
Anaconda, Git, Gerrit, Perforce, Pycharm, CVS, Jira, Confluence
Google Coral TPU, Lattice ECP5, U+, Crosslin-NX FPGA, Raspberry Pi, Intel Movidius, NVIDIA GPU
Compute Engine, App Engine, Vision API, Auto-ML, Container Registry, Kubernetes Engine
Sagemaker, DeepLens, Lambda, Rekognition API, Reko API custom labels
Docker, OpenCV, Machine Learning, Deep Learning, Computer Vision, Convolution Neural Nets (CNN), LSTM, Networking, Model Optimization, Quantization, Pruning, Linux Kernel, OpenWRT
Self-Employed • February 2021 - Present
Working with companies to blend AI with embedded systems specifically to enable AI on edge devices, including the device ecosystem.
Softnautics • September 2016 - February 2021
Sibridge Technologies • May 2015 - August 2016
metric (ALM) as the default metric for path selection.
The project involves enhancing the existing HWMP routing protocol for more efficient working in different environmental conditions and considering other important wireless parameters other than ALM in link cost calculation for better path selection.
Add support for multiple Mesh Points with different channels MIMC (Multi Mesh Interface Multi Channel) for better n/w connectivity and performance by avoiding issues of interference due to the same channel in SISC (Single Mesh Interface Single Channel).
Define both user interfaces of command line and GUI for individual
and central management of the Mesh network
All implementations are on the Linux-based open source code of 802.11s
Development includes understanding of mac80211, nl80211, and cfg80211 drivers as well as utilities like iw, iwconfig, ifconfig, and iwlist.
Integrate power-saving mechanism for multi-radio support in
Linux kernel.
Developed face recognition model compatible with Lattice ECP5 FPGA
Cleaned VGGFace2 with the help of dlib to remove images that could confuse our network
The trained model with the VGGFace2 dataset and custom-added images to give a 128 feature map that can be used to recognize a person’s face
AWS DeepLens