Christen Millerdurai, Developer in Saarbrücken, Saarland, Germany
Christen is available for hire
Hire Christen

Christen Millerdurai

Verified Expert  in Engineering

Machine Learning Engineer and Developer

Saarbrücken, Saarland, Germany

Toptal member since November 9, 2022

Bio

Christen is a deep learning engineer specializing in computer vision, 3D graphics, and NLP. He has expertise in research, development, and deployment pipelines of Deep Learning Products in computer vision and text. Furthermore, his knowledge of DevOps engineering and full-stack makes him an ideal candidate for end-to-end product research, development, and deployment. Christen is highly motivated and focused on achieving goals.

Portfolio

Max Planck Society
C++, NVIDIA CUDA, Artificial Intelligence (AI), Neural Networks, Python, Conda...
starryai
Computer Vision, Generative Pre-trained Transformers (GPT)...
Signzy
Computer Vision, Generative Pre-trained Transformers (GPT)...

Experience

  • Computer Vision - 6 years
  • C++ - 6 years
  • Machine Vision - 6 years
  • Artificial Intelligence (AI) - 5 years
  • Python - 5 years
  • REST - 4 years
  • Natural Language Processing (NLP) - 4 years
  • Generative Pre-trained Transformers (GPT) - 4 years

Availability

Part-time

Preferred Environment

Linux, Git, Docker, Conda, Python, C++, NVIDIA CUDA, Unity, Programming

The most amazing...

...thing I've developed is a solution for extracting information from photos, which is used by customers with over one million daily REST calls.

Work Experience

Research Assistant

2021 - PRESENT
Max Planck Society
  • Researched and developed denoising point clouds using spectral analysis and sampling patterns.
  • Conducted a multi-view mesh parameterization of hands and handled a monocular interaction of two hands using different input modalities.
  • Contributed to multi-view HDR photography for outdoor scene relighting.
Technologies: C++, NVIDIA CUDA, Artificial Intelligence (AI), Neural Networks, Python, Conda, 3D Graphics, Computer Vision, Algorithms, Point Clouds, Discrete Mathematics, Programming

Machine Learning Engineer | Part-time Contractor

2022 - 2022
starryai
  • Contributed to denoising diffusion models to generate artistic images.
  • Halved the processing time of each request, increasing the total revenue by 40%.
  • Provided insights on generating artistic images and tweaked the current generation pipeline to achieve more visually appealing results.
Technologies: Computer Vision, Natural Language Processing (NLP), Generative Pre-trained Transformers (GPT), Neural Networks, Python, Research, Amazon Web Services (AWS), Diffusion Models, Programming

Machine Learning Engineer II

2018 - 2021
Signzy
  • Handled E2E development of a multi-orientation nature scene pipeline for recognizing and extracting text from PDFs and images of ID cards, documents, and payslips.
  • Created an automated machine learning platform for B2B customers that enabled them to annotate, train, and deploy AI models.
  • Built an E2E face authentication and search system for an enterprise with a scalable search feature used by around one million users.
  • Developed and implemented a highly scalable GPU batching interface for Python and C++ to reduce the time needed to create new APIs and increase the throughput for compute-intensive APIs.
  • Led the E2E development of an MRZ extraction solution designed to automatically detect and extract MRZ data from images of nature scenes.
  • Researched text detection, developed a text detection neural network, and published a paper on it.
  • Implemented document quality and image quality analyses to ensure that documents and ID cards were legible and met KYC standards.
  • Contributed to ID card extraction, classification, and cropping solution and worked on document forgery, liveness detection, and mobile ID extraction.
Technologies: Computer Vision, Generative Pre-trained Transformers (GPT), Natural Language Processing (NLP), REST, NVIDIA CUDA, C++, Artificial Intelligence (AI), Neural Networks, Python, Research, Front-end, JavaScript, Vue, Django, Docker, Conda, Kubernetes, Amazon Web Services (AWS), Google Cloud Platform (GCP), Programming

Application Engineer Intern

2018 - 2018
Amazon India
  • Developed a statistical model to analyze customer and service tickets and reduce the number of generated tickets using text mining, NLP, and web scraping.
  • Tracked and fixed payment, authentication, and authorization bugs in the Amazon Appstore.
  • Created internal tools for automatically processing tickets and exporting results with a customizable UI using Python and Django.
Technologies: Python, Django, Natural Language Processing (NLP), Generative Pre-trained Transformers (GPT), Programming

Experience

Open Eye Detector

https://github.com/Chris10M/open-eye-detector
Used Python to create an app that detects whether someone's eyes are open or closed. This software is part of a face authentication pipeline, where face recognition is only done if the person's eyes are open.

Show Me Your Face, and I'll Tell You How You Speak

https://paperswithcode.com/paper/show-me-your-face-and-i-ll-tell-you-how-you
Created a pipeline for lip reading from videos by utilizing lip-to-speech synthesis principles. The pipeline was built using Python and PyTorch and can generate speech for lip-reading content based on the user's facial features.

Car Crash

https://www.youtube.com/watch?v=4Y6rBioJJ9c
Developed a PC game inspired by Temple Run and Subway Surfers using C++. I didn't use any external libraries, only the inbuilt graphics. The challenging parts of the project were handling sound design, using the orthographic 3D camera, and creating assets from scratch.

J.A.R.V.I.S

https://www.youtube.com/watch?v=osWPg8Icmks
Made a speech-controlled bipedal robot that can fit into tight spaces and alert users to the presence of toxic gases by talking or messaging. In addition, I added a face recognition module to the bot.

Kinect-based Motion Detection

Inspired by the famous Tamil movie Jeans, I created this project to thank my teachers for their continuous support. I used Kinect and Unity to construct a visual model of a choreographer and an animated model to mimic his moves in real time.

Natural Scene Text Detection

https://github.com/Chris10M/RFB-Text-Detection
Collaborated on a research project focused on text detection in nature scenes and cropping text in images taken by a phone or web camera. By working on this project, we wanted to improve the state-of-the-art results of ICDAR 2015.

Path Tracer

Built a C++ ray and path tracer from scratch designed for rendering a photorealistic classroom scene. I used OpenMP to accelerate the rendering time and PBRT to implement materials and the model so that it looks photorealistic.

Real-time Semantic Segmentation

https://github.com/Chris10M/Real-time-Semantic-Segmentation
Researched semantic segmentation and developed a pipeline that performed segmentation on 54.5 FPS images in real time. This process included extensive component ablation and used the mIoU value for cityscapes. The project code and model are available on GitHub.

Education

2020 - 2022

Master's Degree in Visual Computing

Universität des Saarlandes - Saarbrücken, Germany

2014 - 2018

Bachelor's Degree in Electronics and Communications

St Joseph's Institute Of Technology - Chennai, India

Certifications

MARCH 2017 - PRESENT

Machine Learning

Stanford University | via Coursera

Skills

Libraries/APIs

PyTorch, Vue, TensorFlow, OpenCV, Keras

Tools

Git

Languages

Python, C++, Embedded C, Embedded C++, JavaScript, C#

Paradigms

REST

Frameworks

Unity, Django, Microsoft Kinect

Platforms

Linux, Docker, NVIDIA CUDA, Raspberry Pi, Kubernetes, Amazon Web Services (AWS), Google Cloud Platform (GCP), Arduino

Other

Computer Graphics, Computer Vision, Machine Vision, Artificial Intelligence (AI), Programming, Natural Language Processing (NLP), Fine-tuning, Data Inference, Generative Pre-trained Transformers (GPT), Conda, Virtual Reality (VR), Geometric Modeling, Speech Synthesis, Robotics, Robot Operating System (ROS), Neural Networks, Research, Front-end, Deep Learning, Machine Learning, 3D Graphics, Algorithms, Point Clouds, Discrete Mathematics, Diffusion Models, Augmented Reality (AR)

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