AuraML offers a cutting-edge Synthetic Image Data Cloud Platform tailored for Enterprises. Eliminating the cumbersome task of manually collecting and labelling training datasets, our solution empowers companies to swiftly generate synthetic data aligned with their specifications. By leveraging AuraML, organizations can expedite the process and reduce costs associated with AI-ML model development, freeing up valuable time and resources for their engineering teams.We are focusing on industries such as Warehouse, Logistics, and Robotics.Work:- Built a scalable cloud platform to generate synthetic image datasets along with annotations. - Supported Annotations Formats: COCO, YOLO - Output Types: Instance Segmentation, Semantic Segmentation, Depth Maps, Normal Maps.- Built a Generative AI-based Data Augmentation Framework that can help users augment their existing dataset by multifolds. - Built an AutoLabelling tool to annotate users' existing real datasets.- Build an LLM-based procedural environment building system where users can use prompts to build large-scale 3d environments. - Identified the memory leaks in the Unreal rendering pipeline and fixed the garbage collection process.Roles and Responsibilities:- Hiring Founding Engineers- Product Management- Team Management- Product Design- Product Demos- Product Documentation- Code Reviews- Synthetic Data Research- Customer feedback and improvement- Built MVP- Finding Customers- Fund Raising- Office Setup in BangaloreTech Stack:- Cloud Platform: Django, ReactJS, PostgreSQL- Core: Python, C++- Rendering: Unreal Engine 5- AWS(EC2, S3, SQS, Lambda)- GCP(Storage Buckets, Compute Instance, Cloud Function, Pub-Sub)

An intensive 6-week programme to nurture and empower early-stage machine learning (AI-first or ML-core) startups. A cohort of top 30 selected startups will get access to high-quality mentorship, curated resources and perks (including AWS credits up to $100k), a community of leading AI/ML startup founders, exclusive passes for in-person events; and an opportunity to pitch to top-tier VC funds, angels, and industry leaders to accelerate their build, growth, and scale-up journey.

Antler is a global early-stage venture capital firm investing in tomorrow's defining technology companies.Highly talented individuals from business and technical backgrounds were part of this AIR1 Cohort - all with the aim to find a co-founder and build a startup. After spending weeks together and knowing each other, we shortlisted the people whom we would like to work with on our startup ideas as co-founders. I was extremely lucky to meet my co-founder in this cohort and start the journey for our startup AuraML together from there.

EF invests in high-potential individuals to help them meet their co-founders, develop their ideas and secure funding from leading investors in the shortest possible time.Worked on 3 different startup ideas:1. Virtual World Platform for hosting social events- host social events such as demo days where users can participate in presentations, sessions, networking, and be part of other social activities that online meeting platforms such as google meet and Zoom do not provide.- Did talk to major accelerators, incubators, VCs, and angler investors for customer development and to know their current experiences, particularly for hosting demo days on our platform.- Paying capacity for this particular use case was not big enough.2. Brain Controlled Wearable Exoskeletons- Brain-controlled exoskeletons can provide better interaction of the exoskeleton with the body parts since the brain has a holistic view of the action that needs to be taken. Currently, exoskeletons are purely mechanical or muscle signals controlled.- Market: Rehabilitation and Defence industry, and for auto-factory workers doing the heavy lifting.- The solution we were providing was too complex and at a higher price point, which didn't make sense to our users.3. Technical Documentation Assistant: - a platform to assist developers and fast-growing tech companies maintain their technical documentation and keep it current.- save time and resources when someone new is joining the company and onboarding them easily- save time during the knowledge transfers when someone is leaving the company- reducing the dependencies on core engineers with time crush, since the knowledge base is maintained and accessible by everyone.- Did customer development with more than 60 companies at different stages.- Developed VSCode Auto Documentation Extension: SyncoAI- Customers didn't find a very big value shift compared to the existing significant products in the market like Confluence and Notion.

Lead and Manage Backend Development Team- Architected and Implemented A/B Test Framework- Managed and Mentored BB Android App Team for Backend Development- Hiring, Training and Mentoring Software Development Engineers - Managed the complete AWS to GCP Migration Process

Responsible for leading the complete Backend Development Process.Using Django for the Josh Skills Mobile Application using PostgreSQL as Database on AWS RDS, FCM for notification, Redis for caching and RabbitMq for Message Queuing, Celery for handling asynchronous tasks, and Firestore for Realtime DB and asynchronous communication, and GCP and AWS as cloud platforms.Major Works:- Mentored Real-Time Multiplayer Quiz Game Development Team- Completely removed Cross-DB Connection from the Django Application.- Managed and Developed Live Conversation Rooms (similar to Club House) within a deadline of 3 weeks.- Implemented Trie Tree Data Structure to reduce user name search time.- Used Redis for Caching and optimized API response time.- Optimized Django Queries using select_related and prefetch_related queryset's methods.- Optimised AWS costs which include the cost of EC2 and RDS.- Upgraded and Migrated complete application from Django 2.3 to Django 3.1.- Implemented peer to peer in-app VOIP call functionality using Agora IO.- Developed complete Points, Awards and Leaderboard system(using Redis Sorted Set for optimization).- Developed Feedback/Grading System for English Reading Practice using SpeechToText Machine Learning Service.- Merging complete data for two users linked through different accounts, including going through every table to dump/resolve conflicting data.- Mentored interns for the Development of Course Content Upload Dashboard.- Serving Compressed images and Videos to the mobile application.- Implemented Round Robin using a linked list for A/B test different versions of the application.- Developed Advanced Admin Dashboards for seamless data management.- Scheduling Tasks such as sending class content, notifications and messages to the user on the specified date and time.- Developed a Subscription-Based Model for selling courses.- Interviewed more than 100 candidates for Software Engineer Role.

Open source contributions to Apache Drill - It is a distributed MPP query layer for self-describing data.Programming Language: Java and C++.Driver: ODBC and JDBCContributions:- Added support of IPV6 address (C++)- Fixed TABLE_TYPE return type issue within DatabaseMetaData.getTables method of JDBC API (Java)- added support for handling complex types (C++)Repository Link: https://github.com/apache/drillPR Links: - https://github.com/apache/drill/pull/1857- https://github.com/apache/drill/pull/1904- https://github.com/apache/drill/pull/1814

Responsible for the development of Apache Drill ODBC and JDBC Data Connectivity Drivers which includes maintenance, bug fixing, improvements, upgradations to the latest standards, and releasing it to our customers. The technology stack used is C++ and Java. Work includes the building of dynamic and static libraries for all platforms (Windows, Linux and macOS). Experienced to work on Visual Studio for C++ Development and Eclipse for Java Development. Salesforce for customer support and Jira for task management.Also involved in contributing to Open Source Apache Drill on GitHub.Software/Tools used:• Visual Studio 2017 for C++ Development• Java Eclipse for Java Development• IntelliJ for Java Development• Tableau for Data Visualisation and ODBC Driver Testing• Microsoft PowerBI for Data Visualisation and ODBC Driver Testing• ODBC Test Tools ODBC Driver Testing• Qlik Sense for Data Visualisation and ODBC Driver Testing• MicroStrategy Desktop for Data Visualisation and ODBC Driver Testing• JIRA for Task Management• Bamboo for CI and CD• Salesforce for Customer Support from Developers• SQuirreL SQL Client for JDBC Driver Testing• Perforce P4V for Version Control• Confluence for Project Documentation

Responsible for the development of Stripe Memphis Data Connectivity Driver. Involved in schema designing to make the full utilization of Stipe Developer APIs to make the connection, query the data and make any required modifications. Then, transforming that schema into predefined JSON format to use it as input for the ODBC and JDBC Data Connectivity Driver.

Developed robust backend and frontend architecture working with Tika Data for auto-labelling images used for computer vision tasks.Automatically scraped images from publically available images, auto-assigned human annotators to annotate the images, and then advanced ML system was able to learn and auto annotate images learning from existing labelled images.

Responsible for Backend Development of 5 websites using Django (Python Web Development framework) as the backend technology stack and using Digital Ocean as the production server. And developing a Hybrid Mobile Application (supported platforms - Android, Windows and IOS) for our customer using Angular JS and Ionic JS.