Ai Game Engineer at Seedify 🚀| MS UMN Twin Cities 👨‍🎓| IIT Kharagpur 🎓| Unity & Unreal C++ Developer 👨‍💻| Ai, Web3 Engineer

Subho Saha

🌟 Senior Game Engineer with 5+ years of experience in software development and 3+ years of excellent US work experience. Building Games and Metaverse one at a time.  🚀

📌 Academics: Completed MS degree from Top 25 university of US. Was admitted to IIT Kharagpur (Top 10 university of India), securing a percentile score of 99.2 in the IITJEE exam.

📌 Engineering Accomplishments: Built Metaverse from scratch with 10+ essential features that facilitated e-commerce. Developed multiplayer event modules for a web3 game, which is serving 30k daily active users.

📌 Skills: Finished 12+ projects across Unity & Unreal, spanning Virtual Reality (VR), Multiplayer, AAA, Generative Ai, Chatbots and Web3 games. Proficient in C++ & C# (6+ years), Multiplayer, and Ai SDKs (Inworld, Convai).

📌 Research: Published 5 research papers, with 100+ citations in Google Scholar  🏅

🔷 GitHub Portfolio: https://github.com/saha0073
_____________________________________________________________________________________________

💪 Core Proficiencies 
🔷 Technologies: Virtual Reality, Metaverse, Multiplayer games, Web3, Machine Learning
🔷 GameDev: C#, C++, Unity3D, Unreal Engine 5, XR SDKs, Multiplayer, Ai SDKs (Inworld, Convai), Blueprints
🔷 Machine Learning: Python, Tensorflow, Pytorch, OpenCV, NLP  
🔷 Others: AWS, SQL, Docker, Flask, React, Oculus, HTC Vive


Creating a cutting-edge AAA Gaming Metaverse with Unreal Engine at Seedify. My role primarily focuses on enhancing the user experience by integrating advanced Ai features.

Checkout the trailer: https://twitter.com/SeedworldMeta/status/1684564606658953218

Ai Game Engineer (Unreal Engine)

Building Kippoverse -> Multiplayer Gaming Metaverse, live in the App Store and Play Store with over 1 million downloads

● Developing real-time gaming metaverse, serving 30k daily active users, resulting in 1 million downloads.
● Created a new Multiplayer event module that allows users to start instant real-time events (like Twitter Space) in the virtual land. 
● Reduced camera jittering by implementing linear interpolation technique and fixing player movement. Finished 20+ bug fixes related to NFTs, UI, colliders, and virtual lands.
● Increased FPS of the app, by identifying O(NxN) RPC calls, and optimizing colliders.

Lead Game Developer

Project 1: Multiplayer e-commerce Metaverse, Nftically (https://www.comearth.world/)

● Developed multiplayer metaverse as SAAS model. Delivered to clients and released it for users to experience the metaverse.
● Worked in coordination with the design team to produce 5 playable prototypes of the Metaverse gallery.
● Conducted interviews, and led a team of 4 people in creating the metaverse.
● Integrated 8 crucial features of metaverse within 3 months, including multiplayer, NFT loading, Ready Player Me avatar, chat, and video streaming.


Project 2: Immersive training simulations in Virtual Reality, IIT Madras

● Completed 3 VR medical training simulations for childbirth using Oculus toolkit, built
controller interaction training, and instrument interaction training simulations.
● Implemented Ai to develop effective training procedures for differentiating experts from novices, trained neural network, and designed subsequent scenes.
● Built immersive AR-based teaching app, in 4 different themes, for the purpose of teaching biology to
schoolchildren using ARCore, and Vuforia.

Metaverse Developer ( Unreal Engine / Unity )

Intelligent Video Analytics Platform for Smart City applications

● Built image preprocessing pipelines, configured Nvidia transfer learning toolkit (TLT), created video analytic pipelines with Deepstream, integrated backend app with web interface.
● Collected and labeled images, created layout in OpenCV, trained object detection, and image segmentation models with SSD, Yolov4, and Mask-RCNN.
● Developed models for 6+ use cases including vehicle tracking, no-helmet detection, and speed violation detection, with transfer learning, increased precision (mAP) up to 80% for all use cases.
● Improved accuracy up to 75% in parking-spot identification, deployed and presented the solutions to the clients, and supported other engineers.

Machine Learning Engineer (Ai)

Development of AI application for automatic fault detection & failure analysis in Disk Drive

● Collected 5k disk-drive images, pre-processed with OpenCV, and built the model pipeline.
● Eliminated the outliers, ran the exploratory data analysis, and calculated decision boundaries to benchmark the model performance and explain critical features. 
● Implemented CNN architecture using Tensorflow, achieved an accuracy of 95% in semantic segmentation.
● Built the backend of the Computer Vision app using Flask & MongoDB, completed user dashboard with React.js, deployed in AWS.
● Increased the classification accuracy from 40% up to 70% through the feature selection
technique and supported factory engineers.

Data Scientist (Ai)

Personalized skill-career recommendations with Ai

SkillEx (Homepage: https://www.skillxploration.com/) is an online tool that helps you plan and achieve your career goals with our unique AI-based training recommendations. I built the NLP recommender, backend architecture, and managed events. You can try our beta here: https://www.skillxploration.com/auth/login 

Achievements: 
● Selected for AWS’s EdTech Startup Accelerator 2021 
● Top Finalist in the 2021 Startup Expo 3.0 by IIT Roorkee
● DPIIT recognition by the Government of India

Co-Founder & Technical Lead (Ai)

Implemented Crystal Graph Convolutional Neural Network (CGCNN) for accurate prediction of material properties of solid-state systems, and boosted performance & accuracy over the existing method.

Machine Learning Research Assistant

• Wrote weekly quizzes, posted solutions online, graded quizzes and exams, kept a record of the scores using Excel.
• Collaborated with professors and other TA’s on the course material, grading policies, and laboratory protocols. 
• Supervised students in problem-solving and conducted tutor hours to help students with homework problems.