AGNETIX - World's most powerful and efficient horticulture lighting

Agnetix is a client based in California. I was the lead Software Developer for developing the web application. The web application controls the light and room parameters. I had the responsibility of architecting and developing the complete UI of the desktop application in a short span of 4 months. I designed several mock ups for various pages of the application. The complexity of constantly maintaining the communication between software & hardware and appropriately modifying the backend codebase made the task more challenging. The technology stack includes React.js in front end and Node.js, MongoDB and InfluxDB in back end. It also contains hardware (light) and to constantly maintain communication between front end, back end and hardware was the most challenging task in the project.

It's mission is to optimize horticulture growth by providing the most efficient and intuitive horticultural lighting solution in the world.

Human Machine Interface

• Timeline: December - May 2019
• Client: Agnetix
• Mentors: Dr. Ihor Lys

BACKGROUND

The Agnetix Human Machine Interface (HMI) is included with the A3 lighting system and delivers visibility into a grow room’s health, providing growers with critical environmental and plant data.

Heat Recovery Technology

The Agnetix A3 platform significantly drives down operational costs with its unique liquid-cooling and heat recovery system. The system efficiently removes heat from the lights using a simple waterloop. Piped out of the room, the heat can be rejected or easily reused to warm the rest of the facility. This results in a more stable environment for better plant comfort, reduced costs and better yields.

Environmental Sensor Series

Agnetix all-in-one, environmental sensors connect very easily to the A3 and offer high-resolution, horticulture readings of temperature, humidity, airflow, CO2, leaf temp and more for precisely monitoring a grow cycle

Close Project

bryte - Web application

I am the lead developer for developing the front end of Admin controlled web application. From the application, an admin can configure the default configurations, add/delete organization, add/delete property and bed devices. Admin can totally control the overall devices from this app.

• Timeline: December 2019 - Present
• Client: Bryte BED
• Mentor: Mark Handel

BACKGROUND

There is a login page, where the admin can login and configure the settings. The admin can manage various organizations and suborganizations. Admin can also change the branding logo.

changing default configuration

Admin can change default configuration from the app.

Organization map

Admin can add/delete/rename organizations from organization map page.

Organization Editing

Admin can add/delete/rename organizations from inside the organization

Bed details

Admin can view bed details from here.

Close Project

Feel The Race - 3-D car racing game

'Feel the Race' consists of making an AI (Artificial Intelligence) based 3D racing game for various platforms with Unity being the primary tool for development. The main aim of the project is to give the gamers a feel of Indian terrains and environment like roads with potholes, bumpers, and rash driving by other drivers which has not been done before. The work includes implementation of various UI and non-UI features and deployment of the game on various platforms like Android and iOS. Thorough research and use of novel algorithms were required for some tasks. I was the lead software developer on Feel The Race and built the game from scratch. My role was to implement novel algorithms and develop racing features (movements, turning, brakes, collision, etc) and real-time graphics along with UI integration in Unity. The project underwent rigorous testing on iOS and Android platforms. Along with the development of the game with different features, I worked on the integration of UI elements in Unity, implementation of Ghost Race, deploying and testing the project on Android and iOS devices.

• Timeline: January - April 2018
• Client: Infocusp Innovations Pvt. Ltd.
• Mentors: Nisarg Vyas (CEO)

Overview of the Project

The project is broadly divided in two parts : UI (User Interface): This includes Introduction screens, Menu screen, Touch Inputs, Map of race track,etc. All the objects with which any user or player spectates, visualises and interacts is covered under UI umbrella. Controller: This includes the functioning of Car - car movement, car steering (left/right), rotation of wheels, brakes, collision of car, etc.

IMPLEMENTATION OF GAME SYSTEM

• Unity Game Engine: The 3D environment of the project is built using the Unity game engine developed by Unity Technologies. Unity is a game development engine combined with an integrated development environment (IDE) that is capable of creating various applications for Android, Mac OS X, Windows, Linux, iOS and other platforms. The Unity game engine can perform physics calculations and graphical rendering, thereby enabling game developers to focus on the development of the game software. Using Unity engine, developers can position, scale, and add realistic physics to objects in the game system.
• Programming language: The primary programming language used for writing scripts to control the behaviors of objects is C#. C# is object-oriented language. It has a variety of libraries and frameworks built specifically for development with Unity, which contains many generic classes and interfaces.

User control settings

Research on Race track

The factors that play an important role in track designing are as follows:
1) Clipping Point: It is a target on the race track edges which maximizes the chances of shortest possible route through the corners.
2) Race track Width: Width of the road is inversely proportional to the difficulty of the player. Width of the road is kept more for a long route, while width of the shortcut for the same route is kept small.
3) Curvature of turns: Angles between the turns should be chosen carefully. Difficulty of a turn increases if the angle between the turn is more acute.

Different modes of game

Future Work

1) Development of AI (Artificial Intelligence) cars which will compete with the player in Quick Race.
2) Integration of google sign in for android devices and apple sign in for iOS devices.
3) Integration of Multiplayer with the game.
4) Improving the existing algorithms.

Close Project

Error Budgets and Developer Insight

Error Budgets: Error Budgets are designed to provide transparency into which services are not meeting their desired stability targets. This in turn helps guide investment towards improving stability of a given service. Error Budgets are computed based on a service or infrastructure defining its Service Level Target (SLT) for its uptime. Given a SLT, we can compute number of failures a service is allowed to have within a defined time frame. Stability Engineering built upon concepts of Error Budgets to create a tool for greater transparency and uniform monitoring across many of our services and key infrastructures.

Developer Insight: One stop solution for managers and developers to analyse their respective projects. A way to summarize all the data that sonarCube produced as well as any other information that we want to analyse, different platforms and then aggregate that and then streamline that data in this portal. Make it easy to add any new data. We have projects, deps file, for a given deps file we suck in sub-data from sonarCube and display it in a table by API call.
Both the projects were large-scale-management portal that displays the details of all the products of FactSet and accurately notifies the owner of the project in case of discrepancies. My full-stack responsibilities involved writing optimized APIs to fetch a large amount of data from the server, implementing the back-end using Python & Node.js, developing a user-friendly front-end using Vue.js and finally deploying the project to production for its targeted users.

<

• Timeline: June - December 2018
• Client: FactSet Research Systems Inc.
• Mentors: Raghavendar Nagamandla (Manager)

Error Budgets

Developer Insight

Close Project

LegalSifter

LegalSifter is a platform aiding lawyers and their associates in finding key terms, clauses, and concepts within a legal contract using sophisticated machine learning and natural language processing techniques. My tasks includes developing a Machine learning tool with various functionalities to extract legal sentences and clauses from legal documents using Natural language Processing. Implemented multiple classifiers for many legal concepts. From sifter kickoff to model packaging, I worked closely with product owners and the annotation team who would assist me in understanding the legal concept upon which I would feature engineer the models and achieve at least 93% F1-score. Technologies used: Pandas, sklearn, spaCy, Stanford CoreNLP and Jupyter.

Why Legalsifter?

Using LegalSifter is easier than reviewing a contract on your own.

• Saves 20-80% review time.
• Advice in a minute or two.
• Finds gotcha terms.
• Tells you what's missing.
• Drives compliance and lowers risk.
• Out-of-the-box or configurable to your best practices.

Our Sifters (artificial intelligence trained to read text) review the contract you are negotiating on yours or other party's paper.

PROCESS

• Timeline: August 2019 - Present
• Client: LegalSifter
• Mentors: Dr. Lars Mahler and Dr. Karin Gregory

SIFTERS

WHAT SIFTERS DO?

Close Project

Online Virtual Classroom

Today in the world of digital things, traditional classes are being replaced by the virtual classrooms or digital classrooms. A benefit to taking online courses is that they offer flexibility to the student. This is a great option for those who already have time commitments with family and work. Online classes will mold with your schedule and allow you to log into your online course at a time that works best for you, as opposed to having to attend a lecture at a specific time. My tasks involved designing mockup designs and developing front end using HTML, CSS, JavaScript and establishing communication with backend using PHP.

• Timeline: August 2017 - December 2017
• Mentor: Professor Prasenjit Majumder

Features implemented to push low fidelity to high fidelity model

• Whiteboard: To increase interaction and ease of communication between professors and students.
• Chatbox: To help students discuss their queries and to create normal interaction between students.
• Video Library: To help professor use/open any video during lectures.
• Book Library: To make storage of all documents easy for users.
• Notes: For helping students to take quick notes
• Raise Hand: To make interface easy for students to ask queries
• Help Assistant: To entertain system and interface related doubts.
• Screenshot: To help the student save any particular whiteboard screen.
• Categorization: The courses were categorized in different fields to provide easy searching.

Tools and Technologies used

• Whiteboard Fox API is used for whiteboard implementation.
• Appear.in API used for video conferencing.
• The server side scripting language used is PHP.
• Software tool used to handle administration of MySQL is phpMyAdmin.
• Languages used for coding web pages are HTML, CSS, JavaScript.
• Chat API used of chatovod.com.
• Help Assistant: To entertain system and interface related doubts.
• Atom used for integration of all the pages.
• Sketch app used for editing pictures.
• Balsamiq tool used for making mockups.

All the information regarding this project can be found here