Sina Afrooze
Founder And CEO at Apera AI
Vancouver, British Columbia, Canada
Summary
👤
Senior🎓
Top SchoolSina Afrooze is a founder and CEO with 10+ years building high-impact AI and robotics products, currently leading Apera AI to commercialize its award-winning Apera 4D Vision™ for factory automation. He combines deep research chops in 3D vision, deep learning, and reinforcement learning with hands-on engineering experience from roles at AWS (Polly TTS and SageMaker RL) and founding work that spun out of his consulting at Creekside Labs. As an early Avigilon engineering leader he scaled R&D to 50 people, helped drive the company to IPO and a $1.2B acquisition, and holds multiple patents in imaging and compression. His open-source contributions include RL-focused work on IntelLabs/coach and practical improvements to Apache MXNet examples, reflecting comfort across frameworks and productionization (ONNX export, multi-framework support, arbitrary-dimensional inputs). Based in Vancouver, he bridges lab-grade research with go-to-market execution, and has a track record of turning “impossible” vision tasks—like grasping transparent parts—into deployable systems.
10 years of coding experience
14 years of employment as a software developer
Bachelors Engineering, Bachelors Engineering at Simon Fraser University
English, Persian
Stackoverflow
980reputation
40kreached
39answers
0questions
Github Skills (18)
python10
mxnet10
machine-learning10
reinforcement-learning10
deeplearning-ai10
deep-learning10
tensorflow10
autoencoder9
onnx9
computer-vision9
nlp8
ctc8
keras6
h2o6
gpu6
Programming languages (3)
Github contributions (5)
apache/mxnet
Dec 2017 - Jan 2019
Lightweight, Portable, Flexible Distributed/Mobile Deep Learning with Dynamic, Mutation-aware Dataflow Dep Scheduler; for Python, R, Julia, Scala, Go, Javascript and more
Role in this project:
ML Engineer
ML EngineerContributions:13 commits, 17 PRs, 136 comments in 1 year 1 month
Contributions summary:Sina's commits primarily focused on improving and extending example code related to deep learning within the MXNet framework. The contributions involved fixing bugs, updating code to be compatible with new MXNet versions, and improving code organization and documentation. Specifically, changes were made to examples relating to NCE loss, autoencoders, and CTC, demonstrating an understanding of various deep learning architectures and training techniques.
pythonschedulerdataflowmutationdata-science
IntelLabs/coach
Oct 2018 - Dec 2018
Reinforcement Learning Coach by Intel AI Lab enables easy experimentation with state of the art Reinforcement Learning algorithms
Role in this project:ML Engineer
ML EngineerContributions:14 commits, 2 PRs, 10 comments in 1 month
Contributions summary:Sina primarily contributes to the reinforcement learning (RL) aspects of the project, as demonstrated by the creation and modification of various head parameter classes (PPOHead, VHead, etc.) within the `rl_coach/architectures` directory, indicating work on defining and configuring neural network architectures specific to RL algorithms. The user also added support for switching between Tensorflow and MXNet frameworks for the project, alongside implementation of a checkpointing framework, with the addition of ONNX export support, illustrating the user's understanding of deep learning frameworks. Additionally, the user introduced the ability to handle arbitrary-dimensional input tensors, expanding the environment compatibility.
starcraftstate-of-the-artdeep-reinforcement-learningstarcraft2carla
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.