Tanvir Sajed
Verified Expert in Engineering
Back-end Developer
Vancouver, Canada
Toptal member since August 30, 2022
Tanvir is a highly skilled software developer and researcher with seven years of experience seeking either Ruby on Rails or machine learning projects. He is a very passionate and meticulous professional, always eager to learn new knowledge and technologies. Tanvir takes ownership, delivers on time, and aims for the best possible outcome in any given scenario.
Portfolio
Experience
- Full-stack Development - 5 years
- Ruby on Rails (RoR) - 5 years
- Machine Learning - 5 years
- Python 3 - 5 years
- SQL - 5 years
- MySQL - 5 years
- PyTorch - 5 years
- Ruby - 5 years
Preferred Environment
Visual Studio Code (VS Code), Microsoft Teams, Linux, Ruby on Rails (RoR), Python 3, Docker
The most amazing...
...tool I've developed is a full-stack ML application presented at the distinguished NLP conference of the Association for Computational Linguistics.
Work Experience
Ruby on Rails Engineer
Self-employed
- Developed, maintained, and updated Phytohub, a plant information related web application using Ruby on Rails.
- Developed back-end data pipelines that import plant-related datasets from online resources.
- Built front-end interface using JavaScript, CoffeeScript, and Slim.
- Worked with Redis, NGINX, Puma, Capistrano, JSON, RSpec, MySQL, Active Admin, and Digital Ocean.
- Deployed the Phytohub application on distributed servers that are being used by researchers at INRAE and other agriculture-based research institutions.
Artificial Intelligence Engineer
Huawei Technologies Co.
- Researched and developed deep reinforcement learning technologies with MCTS, UCB, Bandits, DQN, AlphaZero, AlphaGo, and ResNets.
- Developed ML and RL end-to-end frameworks using Python, C++, TensorFlow, PyTorch, TVM, and Halide.
- Trained very large deep neural networks, including ResNets, using GPU parallelization.
- Deployed a deep RL-based product to GPU, CPU, and NPU servers being used in production.
- Published a paper in ICTAI improving the state-of-the-art results in tensor optimization (TVM) using a Bandit-based reinforcement learning algorithm.
- Published a heuristic search paper in SoCS based on memory-bounded best-first beam search for scheduling halide programs.
Software Developer
Drugbank
- Developed, maintained, and updated a large-scale drug information SaaS Drugbank application with one million daily users using Ruby on Rails.
- Developed back-end data pipelines that import drug-related datasets from online resources.
- Built internal software libraries and RESTful APIs to facilitate biological data exchange using Her and Faraday libraries.
- Worked with Sidekiq, Redis, Elasticsearch, Puma, Unicorn, NGINX, Capistrano, Haml, JSON, RSpec, MySQL, Active Admin, AWS, Google Cloud Engine, New Relic, AppSignal.
- Developed front-end JavaScript and HTML using JQuery, Slim, ERB, and CoffeeScript.
- Implemented SVM, RNN, Transformer, skip-gram, and BOW models for a multi-classification task to predict the type of interactions between drugs and proteins from natural language.
- Implemented spaCy NER classification algorithm to predict different entities like drugs and proteins from the text. Implemented BERT from PyTorch-Transformers for this task.
Artificial Intelligence and Robotics Developer
A&K Robotics
- Researched and developed AI software that predicts objects and their distance from the camera and other sensors like LIDAR and IR, implementing YOLO, R-CNN, and SSD.
- Worked with Python, C++, TensorFlow, PyTorch, Detectron, and Darknet.
- Implemented firmware applications with Arduino and Raspberry Pi and UNIX-based systems on ROS.
- Implemented best practices of Agile software design, Trello cards and epics, pull requests, and GitHub issues.
Software Developer in R&D
The Metabolomics Innovation Center
- Developed, maintained, and updated multiple web application software, servers and databases using Agile software engineering with a team of software developers.
- Created large-scale web applications to handle more than 300,000 users per month on Drugbank and Human Metabolome Database.
- Published five research papers, academic journals, and over ten bioinformatics application servers and databases using Ruby on Rails, MongoDB, MySQL, AngularJS, JQuery, Twitter Bootstrap, and Python Django.
- Developed internal software libraries and RESTful APIs to facilitate biological data exchange.
- Worked on system administration with cloud servers hosted on Google Compute Engine, Digital Ocean, and AWS.
Experience
Stimulating Creativity with FunLines
Based on the generated dataset, we trained a deep neural network model that can predict the degree of fun given a sentence. We then deployed this model in production, which has been used as a guide for users to know how fun their sentences would be. We trained the model using BERT and PyTorch. The application was built using the Python Django framework. The application was demoed and published at the Association for Computational Linguistics (ACL) NLP conference.
Education
Master's Degree in Computer Science
University of Alberta - Edmonton, AB, Canada
Bachelor's Degree in Computer Science
University of Alberta - Edmonton, AB, Canada
Skills
Libraries/APIs
PyTorch, REST APIs, API Development, jQuery, TensorFlow, JSON API, Sidekiq, SpaCy
Tools
Git, NGINX, Capistrano, Google Compute Engine (GCE), SaltStack
Languages
Python 3, Python, SQL, Ruby, HTML, CSS, JavaScript, CoffeeScript, C++11, C++
Frameworks
Ruby on Rails (RoR), Bootstrap, Django, Scrapy, Flask, Darknet, AngularJS
Paradigms
Test-driven Development (TDD), Agile
Platforms
Linux, Docker, DigitalOcean, Google Cloud Platform (GCP), Amazon Web Services (AWS), New Relic, Google Cloud Engine, Arduino, Raspberry Pi
Storage
MySQL, Redis, Elasticsearch, MongoDB, SQLite
Other
Machine Learning, Full-stack Development, Source Code Review, Code Review, Artificial Intelligence (AI), API Integration, Full-stack, CI/CD Pipelines, Back-end, Deep Reinforcement Learning, Deep Learning, Natural Language Processing (NLP), Computer Vision, Expert Systems, Puma, Administration, Heuristics, Monte Carlo Tree Search (MCTS), Residual Neural Networks (ResNets), Beam Search, Genetic Algorithms, JSON REST APIs, Transformers, BERT, Recurrent Neural Networks (RNNs), Support Vector Machines (SVM), Detectron, Deep Neural Networks (DNNs), Robotics, Image Processing, Generative Pre-trained Transformers (GPT)
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring