I, Mithun is a FullStack Software Engineer at Cox Automotive Inc. where I’m currently focusing on building high available, scalable enterprise product which will change the next generation car registration experience for buyers and dealers. We adopted micro-service architecture building on top of battle tested Netflix OSS stack and Amazon Cloud.
I am a technology hobbyist and love to learn and explore new technologies, frameworks, tools. After family and dog, whenever I get free time, I like to code.
I love technology and I love to learn modern & trending technology stacks. Over the years I have learned quite a few technology stacks which I use almost everyday in my personal projects, hobby projects and at my professional workspace. While I love enterprise standard Spring Boot Netflix OSS stack to build micro-service based distributed application architecture, I love the simplicity and flexibility of NodeJS, specially MEAN ( MongoDB, ExpressJS, Angular and NodeJS) stack which gives developer to build entire stack using one common language. Above all I love to build on server-less architecture using AWS Lambda, API Gateway , DynamoDB where ever it makes sense.
When I make fun projects with my daughter or some garage projects or home automation, I like to play with Amazon Alexa, Raspberry Pi, Intel Edison integrating with different AWS server-less services such as SNS, Polly, Lex, IoT
I love to keep generating small project ideas to experiment my new learnings. Over the course, I have executed many personal projects, some are available for public consumption and some are behind home firewall.
11. Car Wallet Pays for Parking Using IOTA (Machine to Machine)
A proof of concept for direct machine to machine micropayments w/o human interaction using distributed ledger-based technology, IOTA Tangle. This project won 3rd price across the globe!
10. DIY Automatic Smart Fish Feeder Using ESP8266 WiFi Module
This is my daughter’s first real project. She is a 3rd grader. Of course, it was a guided project but see learned a lot in the end. 🙂
We have a fish tank with 10+ fish and need to feed them twice a day. It’s very easy to forget as they don’t ask for food like our dog. So we wanted to automate the process and forget about feeding them altogether.
There are many products in less than $20 on Amazon but we decided to do it ourselves and give my daughter some coding exposure. We have used AWS Cloud to update IoT shadow which ESP module listen to. You can use any MQTT broker to establish the communication. We also used ESP deep sleep mode to reduce power consumption. As a downside, it does not run on exact minutes you want but within a window of 60 minutes. It was not critical drawback for us.
9. DIY – Smart Home Surveillance Using AWS Cloud and Raspberry
When I was looking for installing home security camera one of the major concern I had that I don’t trust any vendor to spy on my house through the camera they would install. It’s invasion of privacy and you don’t know if someone is really spying on you. Moreover you need to pay a good amount of month fees for the service. So I decided to build my own. Yeah, it’s not as sophisticated as you can buy in the market but it does the job.
Watch the camera in action in my youtube video below.
Throwing Friday night party at home or hanging out with friends or want to keep your kids busy? Come and Play ‘Tap and Guess’. It’s is a clue based guessing game with loads of riddles, brain teasers , science and trivia questions.
You need Echo Buttons to play this super exciting game.
Buzz fast and score maximum. You get only one chance for each question. If you miss, other player will get a chance to score. There is one golden question in every game for which you score double. Be the first to buzz and beat your friends! There are 3 rounds and 2 questions in each round. Questions get harder as you proceed to higher levels.
6. Blockchain [Live Demo] to streamline vehicle registration process
We at Cox Automotive (Dealertrack, Groton) Fall Hackathon, worked on a concept to streamline vehicle registration process using blockchain.
Vehicle registration is a time taking process where multiple parties are involved and also poses a risk of information manipulation, data duplication and various errors.
Critical information can get highly vulnerable to frauds and data tampering or even become non-traceable.
Bringing Blockchain and moving the entire vehicle registration process on to Blockchain, a lot of these issues can easily be taken care of.
Blockchain, a distributed ledger enables secure recording, management and sharing of information.
Increased transparency and security reduces fraud, waste and abuse.
Interested parties can query latest secure data ensuring they are seeing the latest version of information including title and ownership.
Almost real time transactions saves time and reduce costs.
We built “Auto Metrics” – A system to record the entire vehicle life cycle events in the Blockchain network. We used Ethereum network, Truffle framework, Ganache, IPFS, AngularJS, Web3JS
Watch the youtube video below for full presentation and live demo
5. OfficeBot – A Slack Bot
Project management tools suck. Timesheet is for internal system to track employee’s leaves and typically heavy. There is no simple way to find the whereabouts of a teammate. I personally experienced the need of an easy app where I can query and get the response immediately if someone I am looking for physically present in the office or not. Many organizations are using Slack for collaboration and I thought it would be pretty cool if we can integrate something with Slack and get the things done without wasting money on pricy project management tools and OfficeBot was born!
I built this entire app on server-less technology using aws api gateway, lambda, cloud watch events and dynamodb.
As of Jun 2018, more than 150 teams are actively using this app. Never thought so many people will like the idea. Thanks for using OfficeBot ?
4. Math Challenge- An Alexa Skill
My 2nd grader daughter loves to solve math problems. Often she used to ask me to write down some math problems for her to solve and she wanted me to check her work and provide a grade. I was thinking about automating engagement and at the same time make it more interesting and self-driven. I was reading about amazon alexa skills and was doing some experiment. Suddenly an idea sparked in my brain and wanted to solve my daughter’s problem with the cool voice enabled Echo Dot.
I built Math Challenge using aws lambda and dymanodb.
Quiz Monkey is the most critical and longest personal project I accomplished. I spent almost a year to build the vast and awesome trivia quiz app. Concept, development , content everything has been done by me. This is a hybrid mobile app using Ionic framework and deployed to Google Play Store, Apple App Store and Amazon Play Store. I have learned so much technical stuff while executing this project. Backend was built on top of Loopback (NodeJS) framework, running on multiple cluster node. Realtime communication is achieved by Socket.IO. To be honest, I don’t get time to enhance this project anymore but very proud of what I have achieved to get this product available for general consumption.
Since AWS introduced Rekognition, I have been thinking about a good use case to make my hands dirty with it. I have been doing some Amazon Alexa development for last few months and was wondering about making a Robot with the brain of Alexa which can talk to human. Amazon introduced a python library ASK to run Alexa backend on Raspberry Pi using which you can communicate to Raspberry Pi through Alexa enabled devices such as Echo Dot or Echo Show. But still it was not very pro if you have to integrate other AWS native services.
End of 2016 or early 2017 AWS announced a new service called Lex which provides the advanced deep learning functionalities of automatic speech recognition. Lex uses the same brain as in Alexa and brings deep learning capabilities to developers. Using Lex and Rekognition with the power of Raspberry Pi you can make something very cool and interesting AI projects.
I have created a Robot , named Tintin who is capable of interpreting your speech as commands, take pictures of yours and run it through facial recognition algorithm and greet you with your name. For the very first time when Tintin meets you, he is smart enough to ask your name and memorize it. So upon your next visit to our house, Tintin will recognize your face and greet you with your name.
1. Ethereum Smart Contract with Internet of Things
Recently I have been working on another personal project which I like to share with you. I chose a peer to peer online car sell use case and implemented using Ethereum blockchain network. This is not a real product to sell car online rather just a use case to show how we can leverage blockchain network to solve this kind of use cases.
The concept is very simple. Seller will post an ad online for the car s/he wants to sell providing vehicle make/model/year and listing price. System will create a block in the chain using Ethereum smart contract (contract written in Solidity). Seller can search different vehicles listed for sale and choose a car s/he wants to buy. When buyer confirm the purchase another block will be created in the chain. As you know blocks in the chain are immutable which makes the system secured and hack proof. When purchase is complete, system will generate a QR code. Buyer walks to the garage where car is parked with the QR code and scan it by the web camera installed in front of the garage. Upon successful ownership verification of the QR code against the block chain data, garage door will open. At the same time, system will send a text message to seller providing a link to watch the live streaming of his/her garage. QR code verification and open/close of the garage are achieved using Raspberry Pi 3, Servo Motor and few AWS managed services such as IoT Shadows, SNS and Polly. Please note, this use case is not market ready and need to include authentication, strong verification mechanism, fraud listing etc. These are out of scope of my project as I focused on core technology use case rather than real business use case.
Watch the demo below.
I have used a test ethereum network called TESTRPC rather than using real public Etherem network. TESTPRC is easy to run on local machine ( Windows/Mac/Linux) and does not need real crypto currency to complete a transaction.
“Everything fails all the time” — Werner Vogels This is sad but true, everything fails specially in Microservice architecture with many external dependencies. Modern applications have tens of microservices which communicate with each other over REST. Anytime any microservice may go down causing entire operation to fail. At a broad level we can classify these …
I got very excited when I first saw Microsoft Bot Framework (https://dev.botframework.com). Microsoft has done a great job! You can deploy same bot to multiple channels such as MS Team, Slack, Facebook. But all the documents/tutorial I read recommend to deploy bot backend code to Azure. I have been working with AWS for few years …
If you are a Raspberry Pi developer you often need to access web server running on your Raspberry Pi from outside of your home network. For example on your mobile when you are not connected to home Wifi. There are different ways to let the traffic go to your Raspberry Pi from internet such as …