Apr 2024 – Jul 2024 · 3 mos · Amherst, Massachusetts, United States · Hybrid

• Engineered and integrated Auth0 authentication workflow for a college web application using Azure Entra ID with OIDC protocol enhancing faculty login experience and security.
• Architected a robust system from scratch using React, Next.js, and Django-Ninja to streamline academic data ingestion into a relational data warehouse for multiple clients, leveraging RabbitMQ and Celery for efficient pipeline performance.

Feb 2023 – Nov 2023 · 9 mos · Amherst, Massachusetts, United States

• Worked in a cybersecurity lab a.k.a Rescue Lab, building software for assisting DOJ in efforts to eliminate internet crimes against children.
• Spearheaded the transformation of a web application, driving the evolution from POC to a robust MVP by designing and implementing critical features for FastAPI backend and React frontend microservices.
• Devised a tailored database ORM using Python, Pydantic, and OOP concepts, optimizing communication between FastAPI and Azure Table Storage for enhanced efficiency.

May 2022 – Aug 2022 · 3 mos · Menlo Park, California, United States

• Engineered efficient APIs and Hadoop data pipelines for manufacturing and extracting 24 new features, related to Facebook awards, to potentially improve the existing group feed ranking model.
• Performed feature selection and trained a probabilistic model for getting the likelihood of a post being awarded and achieved a 4% improvement in AUC as compared to the model trained without the newly added features.

Jan 2022 – Jul 2022 · 6 mos · Remote

• Worked on evaluating and assessing current super-resolution methods for both images and videos to assess their performance in terms of latency and super-resolution quality, and researching new strategies to improve the real-time execution of these methods.
• Benchmarked state-of-the-art video super-resolution(VSR) models against various datasets including a novel categorical video test dataset which we contributed for the in-depth evaluation of videos.
• Proposed a design and built a simulation of a new strategic model using Python and Pandas, interleaving Video Super Resolution(VSR) and Video Interpolation(VI), to super-resolve the live video streaming.
• Gained 34% efficiency (Publication 3) over baseline VSR model in output video frames/sec for high-resolution videos.

May 2021 – Aug 2021 · 3 mos · Amherst, Massachusetts, United States

• Built a system to bring up a Django-powered data visualization and geospatial analytics web application, helping local law enforcement in reducing crime-incident tracking efforts by 60%.
• Implemented a robust CI/CD pipeline using Github Workflows, AWS Code Pipeline, and AWS EC2 for automated testing and deployment, enhancing software development efficiency.
• Automated data ingestion process for the client to easily upload new crime data in the system.

Apr 2019 – Nov 2020 · 1 yr 7 mos

• Leveraged agile and test-driven development methodologies to develop Flask Rest APIs for provisioning micro-service, which involved automation for deploying and managing Citrix infra. across cloud environments like AWS and Azure.
• Devised a new architecture to perform autoscaling in distributed cluster systems efficiently and robustly using the concept of the cold node. Achieved 57% time performance improvement in cluster scale-out over previous architecture.
• Revamped the existing workflow to introduce the flexibility of retrying, halting, and resuming cloud cluster provisioning and de-provisioning, and enhance user experience in case of any server/client fault.
• Reduced the overall cloud storage cost for the customers by automating the process of checking out Citrix (CICO) license as part of the infrastructure deployment workflow.

Apr 2018 – Mar 2019 · 11 mos

• Worked on integrating Citrix ADC (Netscaler) products with cloud orchestration platform OpenStack, enabling Citrix ADC’s features to be consumed by the cloud users.
• Worked on architectural changes in Cloud Orchestration services for enhancing efficiency and user-friendliness.
• As a part of adopting new architecture for cloud orchestration services for OpenStack, automated the complete infrastructure and PostgreSQL database migration flow for hassle-free adoption to the new framework.

Jul 2017 – Mar 2018 · 8 mos

May 2016 – Jul 2016 · 2 mos · Varanasi Area, India

• Real Time Text Detection Using Mobile Phones
• Developed an Android-based text searching application capable of recognizing and locating text in real-time camera feeds by showing bounding boxes around the searched text.
• Used OpenCV for image processing, MSER algorithm to detect the region of text and Tesseract OCR to recognize the text.
• Completed under the guidance of Dr. S. K. Singh, Department of Computer Science Engineering, IIT(BHU).

Feb 2016 – Feb 2016 · 0 mo · Varanasi Area, India

• Worked as Coordinator for coding event, Codacity in the EE Department’s annual technical festival, Prastuti’16.
• Incorporated the flavor of Computer Science and Electrical Engineering for framing intuitive and interesting problem statements for the event.
• Conducted a series of workshops to equip participants with the right set of skills for the event.

May 2015 – Jul 2015 · 2 mos · Varanasi Area, India

• Economical Efficient Sentiment Stream Analysis
• Implemented the paper "Economical Efficient Sentiment Stream Analysis" and worked on enhancement of the paper under the guidance of Dr. R. N. Chowdary, Assistant Professor, Department of Computer Science Engineering, IIT (BHU).
• Worked on enhancing selective sampling algorithms, namely Pareto(PSS) and Kaldor(KSS), so as to deliver a system that can perform Sentiment Stream Analysis of Twitter data more efficiently. Proposed algorithm was efficient than KSS but not to PSS.
• Language: Python