Aug 2024 – Present · 1 yr 10 mos · Hybrid

• 1. Developed microservices for an end-to-end portal using a SHAP based residual deep learning model to classify and predict heart diseases based on ECG signals across 51 classes.
• 2. Reviewed and benchmarked existing text-to-speech models and their real-time factors; fine-tuned various opensource TTS models including MelloTTS, CoquiTTS etc over Indian-English accents

Jun 2023 – Jul 2024 · 1 yr 1 mo · Jakarta, Indonesia

• 1. Developed proof-of-concept (POC) solutions for plant disease detection, segmentation, and classification tasks by crafting a customized deep learning architecture. The architecture incorporated the AlexNet framework with ResNet serving as the backbone, thus leveraging the strengths of both architectures. The model was meticulously fine-tuned by applying advanced techniques, including optimizing learning rate schedules to facilitate faster convergence. Other data augmentation techniques like stain normalization, image rotation etc were used to to enhance the model's robustness to variations in image quality and to further capitalize on the potential of transfer learning, the model was integrated with pre-trained weights from state-of-the-art networks.
• 2. Developed a personalized credit scoring system to lend loan as part of Buy Now Pay Later product leveraging the specific attributes and behaviors of users to provide tailored credit or loan decisions
• 3. Complete ownership of over 15 features from the date of joining along with optimizing runtime of various APIs by an astounding 90-94% and reducing warnings in our backend system by over 85%. Go to engineer for performance, efficiency and user-centric optimizations; resolved over 100 high-priority issues and bugs improving overall user experience

May 2022 – May 2023 · 1 yr · Jakarta, Indonesia

• 1. Developed a search engine from scratch using AWS open-search, resulting in a 60% improvement in user retention and 80% improvement in order conversion. It involved creating and maintaining indexes, implementing data synchronization, and crafting optimal elastic-search queries.
• 2. Developed a core service for product recommendation system utilizing content-based filtering for personalized user-user and product-product recommendations.
• 3. Developed a robust, in-house role-based authorization control system for multiple product suites.

Dec 2021 – May 2022 · 5 mos · Remote

• 1. Architecture and Features: Built a cutting-edge Role Based Access Control (RBAC) system using OAuth 2.0 and OpenID Connect. Integrated various Oauth 2.0 solutions like Auth0, Okta, Keycloak, AWS Cognito into the RBAC system.
• Currently the algorithms are used across 4 different products
• 2. Logging: Developed a custom logging library for streamlined logging across multiple product suites.
• 3. Performance and Optimization: Involved in resolving security audits issues by around 80%; reducing warnings and errors in the MVIS product suite by an impressive 90%.

Jul 2020 – Dec 2020 · 5 mos · Wilmington, Delaware, United States

• I worked as a remote SDE intern with JWilliamson INC. I worked on the backend system of certain plugins and developing APIs focussed in JAVA along with GraphQL adapters that exposes the APIs over HTTP and uses certain data fetchers for deserialization and serialization of the data. Other libraries such as Bukkit, Spigot and BungeeCord are integrated with the plugins for seamless cloud based deployment. We follow rapid development using the agile development methodology producing solutions that are innovative and yet robust

Sep 2017 – Nov 2018 · 1 yr 2 mos · New Delhi Area, India