Hakob Kirakosyan, Developer in Yerevan, Armenia
Hakob is available for hire
Hire Hakob

Hakob Kirakosyan

Verified Expert  in Engineering

Bio

Hakob is a machine learning expert and scientific programmer with substantial experience delivering analytics and predictive solutions for business and scientific applications. Additionally, he enjoys developing video games based on augmented reality and remote human-computer interaction paradigms. Hakob is fluent in Python and C++ and has worked as a consultant in a variety of projects, helping clients achieve their goals.

Portfolio

RunKicker Pte Ltd
Artificial Intelligence (AI), Signal Processing, Python, Video Processing...
Strong Analytics, LLC
Computer Vision, OCR, Deep Learning, UPC & Barcodes, TensorFlow, PyTorch...
Max Planck Society
OpenCV, DNN, Deep Neural Networks (DNNs), Open Neural Network Exchange (ONNX)...

Experience

Availability

Part-time

Preferred Environment

Vim Text Editor, Linux, Windows, Visual Studio, PyCharm

The most amazing...

...thing I've done is implement a real-time human segmentation algorithm on mobile platforms along with stunning particle diffusion effects.

Work Experience

AI Expert

2022 - 2023
RunKicker Pte Ltd
  • Developed a deep learning model for BP calculation from PPG signals.
  • Optimized and fixed the heart rate calculation algorithm from PPG signals on Android devices.
  • Implemented the finger guidance algorithm for finger positioning control to collect better PPF signals.
  • Deployed AI models on Android devices and optimized them for high-performance usage.
Technologies: Artificial Intelligence (AI), Signal Processing, Python, Video Processing, Image Processing, Health, Computer Vision, Video Analysis, C++, Models, PyTorch, TensorFlow, Mobile, Signal Analysis

Computer Vision Expert

2021 - 2022
Strong Analytics, LLC
  • Developed a visual symbology image enhancement algorithm, which improved the decoding rate by around 10%.
  • Developed an exceptionally fast barcode detection and rectification method with deep learning.
  • Optimized above-described NNs for edge devices, getting incredible, up to 100 FPS performance on very limited devices.
  • Optimized the plant and stem detection algorithm 10x while improving accuracy.
  • Implemented fast visual odometry working on edge devices for a robot.
Technologies: Computer Vision, OCR, Deep Learning, UPC & Barcodes, TensorFlow, PyTorch, Machine Learning, Data Science, NumPy, Artificial Intelligence (AI), Video Processing, Image Processing, PySpark, Pandas, Scikit-learn, Deep Neural Networks (DNNs), Python, Machine Learning Operations (MLOps), NVIDIA CUDA, You Only Look Once (YOLO)

Senior Computer Vision Engineer

2021 - 2021
Max Planck Society
  • Developed a back end for scientific experiment conduction and data collection.
  • Built a video chatting system between experiment participants.
  • Integrated real-time face detection and random blurring during the video call between participants.
Technologies: OpenCV, DNN, Deep Neural Networks (DNNs), Open Neural Network Exchange (ONNX), Django, MySQL, Machine Learning, NumPy, Artificial Intelligence (AI), Video Processing, Image Processing, Scikit-learn, Computer Vision, Deep Learning, Python, Machine Learning Operations (MLOps)

Senior AI Developer

2020 - 2020
Client (via Toptal)
  • Implemented real-time human pose estimation on mobile devices.
  • Developed and implemented an AI algorithm to detect certain movements of the user from a real-time camera stream.
  • Developed an AI algorithm that provides real-time human form guidance and gives feedback about the repetition of a specific movement.
Technologies: iOS, OpenCV, Swift, Core ML, PyTorch, Python, C, C++, Machine Learning, TensorFlow, NumPy, Artificial Intelligence (AI), Video Processing, Image Processing, Pandas, Scikit-learn, Computer Vision, Deep Neural Networks (DNNs), Deep Learning, Machine Learning Operations (MLOps), NVIDIA CUDA

Senior AI Developer

2019 - 2019
UBOT
  • Built an autonomous control system for citrus harvesting robots.
  • Implemented citrus instance segmentation algorithms and their 3D position estimation through Azure Kinect.
  • Designed and implemented GPS-based autonomous driving logic to drive robots in citrus farms.
Technologies: Microsoft Kinect, Bash, PyTorch, C, C++, Python, Machine Learning, Data Science, NumPy, Artificial Intelligence (AI), Video Processing, Image Processing, Computer Vision, Deep Neural Networks (DNNs), Deep Learning, Machine Learning Operations (MLOps), NVIDIA CUDA

Senior AI Developer

2018 - 2019
Edgecase
  • Developed a hybrid environment to simulate various 3D scenes for synthetic data acquisition.
  • Used synthetically generated images to train deep neural networks for object segmentation.
  • Developed and implemented algorithms based on graph theory for automated 3D environment generation.
  • Designed and developed structured databases for images with corresponding annotations as training data for further deep learning usage.
Technologies: Amazon Web Services (AWS), Google Cloud Platform (GCP), Azure Cloud Services, Unity, C#, C++, Python, Machine Learning, Data Science, TensorFlow, NumPy, Artificial Intelligence (AI), Video Processing, Image Processing, Computer Vision, Deep Neural Networks (DNNs), Deep Learning, NVIDIA CUDA

Senior Computer Vision Engineer

2014 - 2018
Improvis
  • Built and integrated real-time human segmentation algorithms on mobile platforms.
  • Developed face detection and alignment calculation algorithms on mobile platforms.
  • Implemented strawberry detection and quality estimation algorithms using UAVs (unmanned aerial vehicles) video streams.
  • Created and integrated background extraction algorithm for object segmentation and real-time inpainting.
Technologies: OpenGL, C#, Dlib, OpenCV, C, C++, Python, Unity, Machine Learning, Data Science, TensorFlow, NumPy, Artificial Intelligence (AI), Signal Processing, Video Processing, Image Processing, Cryptography, Pandas, Computer Vision, Deep Neural Networks (DNNs), Deep Learning, NVIDIA CUDA

Automated Placement and Routing System for 3D Objects in a Unity Synthetic World

https://www.edgecase.ai/
I developed and implemented algorithms based on graph theory and VLSI design theory to generate variadic synthetic 3D worlds and routing of movable objects to generate large numbers of synthetic data images to increase the accuracy of deep learning models.

GravitX iOS App

https://gravitx.improvis.ai/
I developed an iOS game that allows interacting with game objects using touchless control techniques like head and eye-based manipulation. The AI backbone was based on iPhone/ iPad True Depth sensor data, with additional processing logic to increase sensitivity up to a 1MM level.

Autonomous Harvesting Robot

I created an instance segmentation backbone with 3D position extraction of citrus fruit in the robot gripper's coordinate system. I also designed a transfer learning pipeline to increase the detection system's accuracy by enlarging the training data set.

Hide Magic Mobile App

I developed a Unity-based cross-platform app that allows the automatic segmenting of foreground objects from the video camera stream and hiding/diffusing with real-time background painting. We implemented particle rendering using a Unity Native particle system.
2020 - 2022

Master's Degree in Computer Science

Yerevan State University - Yerevan, Armenia

2015 - 2019

Bachelor's Degree in Informatics and Applied Mathematics

Yerevan State University - Yerevan, Armenia

Libraries/APIs

PyTorch, TensorFlow, OpenCV, Keras, NumPy, Pandas, Scikit-learn, Eigen, Dlib, OpenGL, PySpark

Tools

Visual Studio, Vim Text Editor, Open Neural Network Exchange (ONNX), You Only Look Once (YOLO), Xcode, PyCharm, CMake, MATLAB

Languages

C, Bash, C++11, C++14, C++17, Python, C++, C#, SQL, Swift, JavaScript

Frameworks

Unity3D, Qt, Django, Unity, Qt 5, Core ML, Caffe, .NET, Microsoft Kinect

Paradigms

Concurrent Programming

Platforms

Windows, Amazon Web Services (AWS), Linux, DNN, NVIDIA CUDA, Azure, Docker, iOS, Android, Google Cloud Platform (GCP), Kubernetes, Mobile

Storage

Amazon S3 (AWS S3), Data Pipelines, Azure Cloud Services, MySQL

Other

Image Recognition, Statistics, Unity3D Native Plugin Interface, Algorithms, Data Structures, Neural Networks, Object Detection, Multithreading, OCR, Convolutional Neural Networks (CNNs), Machine Learning, Deep Neural Networks (DNNs), Deep Learning, Graph Theory, Numerical Methods, Probability Theory, Game Theory, Differential Equations, Mathematics, Microsoft Azure Cloud Server, Computer Vision, Image Processing, Artificial Intelligence (AI), Optimization, Computer Vision Algorithms, Data Science, Signal Processing, Video Processing, Linear Regression, Machine Learning Operations (MLOps), Video Analysis, Augmented Reality (AR), Virtual Reality (VR), Generative Pre-trained Transformers (GPT), Generative Pre-trained Transformer 3 (GPT-3), Cryptography, UPC & Barcodes, Health, Models, Signal Analysis

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring