Dhananjoy Biswas, Developer in Dhaka, Dhaka Division, Bangladesh
Dhananjoy is available for hire
Hire Dhananjoy

Dhananjoy Biswas

Verified Expert  in Engineering

Software Developer

Location
Dhaka, Dhaka Division, Bangladesh
Toptal Member Since
June 7, 2019

Coding since 2010, Dhananjoy started working professionally in 2013. After leaving Google in 2015, he joined a YC-backed startup and scaled it from 3 to a 50+ person team to become the largest eCommerce in Bangladesh. From the start, he has always enjoyed taking on new challenges besides his day-to-day works. He is also a 4x IMO (Intl. Mathematical Olympiad) and IOI (Intl. Olympiad in Informatics) medalist and 2x ACM ICPC world finalist.

Portfolio

Virtually
TypeScript, React, Next.js, Firebase, Serverless
Toptal
React, GraphQL, TypeScript
Dryft
Amazon Web Services (AWS), WebSockets, Node.js, AWS Lambda, Video Streaming...

Experience

Availability

Full-time

Preferred Environment

Visual Studio Code (VS Code), Vim Text Editor, Ubuntu, MacOS

The most amazing...

...thing I've built is an eCommerce search engine, serving 30+ million products refreshed every 6 hours and trained using deep neural net to auto-learn ranking.

Work Experience

Senior Full-stack Engineer

2020 - 2021
Virtually
  • Developed and quickly iterated new features to serve the fast-growing number of educators moving classes online during the pandemic, which led to a surge in growth and helped the startup raise $1.75 million in funding.
  • Mentored other engineers by providing feedbacks, code reviews, and through one-on-ones as the only senior engineer.
  • Leveraged SWR (swr.vercel.app) and multi-layer caching to make the web application run blazingly fast and to be offline-friendly.
  • Refactored firebase integration to secure and offload critical and complex transactions from the front end to Firebase cloud functions.
  • Introduced linter, Prettier, and TypeScript to supercharge developers' productivity across the team.
Technologies: TypeScript, React, Next.js, Firebase, Serverless

React Front-end Engineer

2020 - 2020
Toptal
  • Developed new features for the Toptal Talent Portal.
  • Developed new pages and features for Toptal Staff Portal, the internal management system of the organization.
  • Contributed to the internal React component library shared across multiple teams.
Technologies: React, GraphQL, TypeScript

Back-end Software Engineer

2020 - 2020
Dryft
  • Leveraged Twilio Video API and Google Chrome API to record live video from user webcam and stream to the server for processing.
  • Created a serverless solution to process live video streams and generate a YouTube-style live trailer from them.
  • Built a live and DVR-like video delivery service with AWS MediaLive, MediaPackage, and MediaStore.
Technologies: Amazon Web Services (AWS), WebSockets, Node.js, AWS Lambda, Video Streaming, Twilio API

Lead Software Engineer

2020 - 2020
Mark Cuban Cost Plus Drug Company
  • Led a team and actively contributed to building a full-featured eCommerce solution from scratch with React, Node.js, and GraphQL.
  • Implemented automated deployment and a CI/CD pipeline with GitHub actions.
  • Containerized the front-end and back-end services with Docker and deployed the services with Kubernetes.
  • Integrated Stripe for processing payments and Stripe Radar for fraud detection.
Technologies: React, Stripe, Kubernetes, Google Cloud Platform (GCP), Docker, Node.js, GraphQL

Lead Software Engineer

2020 - 2020
Private Govt Agency
  • Led a team to design and build a mobile application with React Native that was used by law enforcement authorities to contain the coronavirus pandemic by tracking and making sure that quarantined people are staying at home.
  • Researched and implemented geofencing solutions to track and monitor smartphone locations.
  • Built a face recognition service able to run on smartphone devices to accurately identify and verify quarantined people are staying home by matching their faces.
Technologies: Algorithms, React, Firebase, React Native, TensorFlow, Python

Founder, Engineer

2015 - 2020
Backpack
  • Actively developed from scratch and later led a team of developers to build an eCommerce website with modern technologies (TypeScript, React, and Node.js) that served 30+ million products to 200,000+ shoppers.
  • Led a team to design and implement a referral program (first of its kind in Bangladesh) that resulted in a 35% average MoM growth to make the company the largest eCommerce in Bangladesh.
  • Created a full-text, image-based product search engine on top of Elasticsearch to serve 30+ million products.
  • Built a learning-to-rank model to learn relevance from user interactions with TensorFlow automatically.
  • Scaled and maintained several Kubernetes clusters consisting of 200+ nodes and 50+ services, deployed in AWS, and then migrated to GCP.
  • Wrote a large-scale crawling infrastructure with headless Chrome and serverless technologies (AWS Lambda, Google Cloud Pub/Sub) to crawl real-time product data and prices from major eCommerce sites including Amazon, eBay, Walmart, and Best Buy.
  • Developed a serverless image processing service with AWS Lambda to generate customized advertising images for tens of millions of products on the fly.
Technologies: Elasticsearch, Machine Learning, React, Google Cloud Platform (GCP), AWS Lambda, Docker, Kubernetes, TensorFlow, Python, Node.js

Software Engineering Intern

2015 - 2015
Google
  • Built a massively distributive predictive model to forecast the rate and frequency of changes in flight prices of hundreds of carriers.
  • Scaled the model to analyze hundreds of terabytes of cache data by parallelizing to 5,000+ nodes.
  • Optimized the price caching algorithm to use 12% less computing resources, thus massively reducing the cost by freeing up ~150,000 servers.
Technologies: Algorithms, R, C++, Python

End-to-End Product Search (TensorFlow/LearningToRank/Elasticsearch)

At my current international peer-to-peer delivery startup, I've built a product search engine which crawls realtime product information from major US-based eCommerce websites and then learns to rank them for search and recommendation.

There are three major components involved:
- A fleet of crawlers built with headless chrome running on a serverless architecture, capable of crawling 10000+ product pages per second
- A highly scalable Elasticsearch server for storing 30+ million product information and hundreds of millions of analytics data, also acting as a basic full-text search engine
- A learning to rank model, built with Tensorflow, which learns ranking from search results from original sites and also from user interactions.

The ranking model consists of several submodels
- a modified word2vec algorithm which learns popularity and contextual embedding of a product
- an ensemble of sentence encoders which learns the semantic embedding of a product from textual data like name, features and descriptions
- an image encoder to compute fixed-sized embedding from product images
- an bidirectional RNN to embed queries into fixed size vectors
- and finally another RNN to embed user interactions to fixed size vectors

Predicting Weight and Dimensions of Product From Image and Description

As an international peer to peer delivery startup, efficiency is everything at my current startup, Backpack. And it always comes down to the last 1%.

We rent traveler's luggage space to carry products from one country to another, and it's paramount to utilize the space as efficiently as possible. Overestimating can lead to excess luggage penalties where underestimating leads to unutilized resources. So we need to know to weight/dimensions of the products we deliver as accurately as possible.

For this, we've built a predictive model which can predict the weight/dimensions of a product from its images and descriptions which enables us to utilize the luggage space with 99.4% accuracy.

Real Time Price and Product Data Scraping API

https://rapidapi.com/ebappa1971/api/amazon-price
A SaaS API to search for products and fetch the real-time price and product data from top eCommerce websites, built on top of serverless architecture using Google Cloud Functions and written with browser automation technologies using Node.js and Puppeteer. The service has 1,000+ monthly active users who are using the API to track competitor prices and market trends.

Languages

SQL, JavaScript, ECMAScript (ES6), Python, TypeScript, C++, C++17, R, GraphQL, Haskell, Java

Frameworks

Express.js, Redux, Serverless Framework, React Native, Next.js

Libraries/APIs

Node.js, React, TensorFlow, Stripe, Twilio API

Tools

BigQuery, Google Kubernetes Engine (GKE), Vim Text Editor

Platforms

Linux, AWS Lambda, Amazon Web Services (AWS), Google Cloud Engine, Kubernetes, Docker, Google Cloud Platform (GCP), Firebase, MacOS, Ubuntu, Visual Studio Code (VS Code), Android

Storage

MySQL, Elasticsearch, Redis, PostgreSQL

Other

Computer Vision, Natural Language Processing (NLP), Deep Learning, Serverless, Data Mining, Machine Learning, Algorithms, Google Cloud Functions, GPT, Generative Pre-trained Transformers (GPT), Data Cleaning, Mathematics, Video Streaming, WebSockets

Paradigms

Data Science, DevOps

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