Parth Soni

Developed and maintained the back-end integration layer for an AI‑powered conversational platform (akin to ChatGPT) using Java microservices to bridge a React‑based UI with Python‑driven LLM services. Architected RESTful APIs with authentication, rate‑limiting, error handling, and retries, and integrated AWS SQS for message queuing, S3 for prompt and log storage, Step Functions for workflow orchestration, and Batch for heavy‑compute tasks. Implemented response parsing and normalization, fed enriched data into a Solr index for fast, contextual search, and built monitoring and logging pipelines to ensure reliability and performance. Collaborated closely with front-end engineers and data scientists to launch new features, drive latency optimizations, and deliver a seamless user experience.

A call control widget integrated with SAP C4C that allows customer service agents to perform call operations, including placing outgoing calls, answering incoming calls, transferring, holding, and conferencing through the web interface instead of using their softphones.

Designed, architected, and developed the application from scratch per discussion with the end client, Prudential. Designed event-driven architecture through WebSockets as per the application's need. Enabled high availability through failover to available secondary servers using service registry, API gateway, and ActiveMQ.

Cloud agnostic event-driven rule engine. A microservices application for GSK that processes illness data from all over the world and combines it with configurable rules that auto activates multiple ad campaigns in specific locations at the best time of the day, month, and year.

I optimized services by adding an external Redis cache service to save the computing units and configured services to run on Docker and Kubernetes cloud environments.

Set up Jenkins for CI/CD and the media campaigns on advertising platforms like Google SA360, DV360, DV360 YouTube, Twitter, Facebook, and TTD. Monitored alerts and daily cloud logs in Google Cloud Platform (GCP).

Identified and processed different data sources to gather and rank campaign-specific keywords using Python scraper and respective APIs.

DAM is a digital asset manager that stores digital assets, including card arts, privacy policies, and the terms and conditions of AEXP's partners. DAM also manages the versions of assets having historical data.

Designed a solution that could be easily integrated into their existing environment. Consolidated with AWS cloud, created APIs to store and retrieve the digital assets, and integrated AEXP's single sign-on for authentication.

TopRecruiter is a recruitment website helping recruiters to find the right candidate. A serverless platform from back-end services using MongoDB cloud, Elasticsearch, and Azure Logic Apps.

I worked on designing the MongoDB database model and developed stitch webhooks to expose data in the required format.

Enabled security of MongoDB stitch third-party services using an API-key provider and built search API from a configured Elasticsearch index.

Set up the local development environment using MongoDB Stitch Command Line Interface (stitch-cli) and configured Jenkins for automatic application deployment and provisioning to different environments.

Radisson Hotels have different brands, such as RadissonBlu, ParkInn, and RadissonCollection. These sites were driving traffic from multiple content pages hosted in Oracle WebCenter Sites.

I designed the Asset APIs and architectural reusable CSElements in Oracle WebCenter Sites and Fatwire.
Developed a new Flex Assets to cache the response of the external requests.

I also developed RESTful web services using JAX-RS and consumed RESTful services that parse XML to Java objects to implement content as a service from a third-party content provider.

Integrated Facebook and Twitter accounts using the Engagor API and designed and improved the caching strategy to remove the cache dependencies for minimum cache flush while publishing.

Ketu is a product that helps pharmacies to provide specialty services and manage workflow. It allows patients to avail different programs to support the cost of their specialty medication.

I was involved in SDLC requirements gathering, analysis, design, development, and testing applications developed using Agile Scrum methodology.

Worked on multitenant applications by integrating the shard grails plugin to switch across multiple database connections. Optimized scheduled cron jobs by enabling multithreading to increase performance and decrease completion time.

I also developed and maintained Unix shell scripts for data-driven automatic processing and used Gradle for building WAR files deployed in application servers for dependency management of code.

Resolved bugs and defects in the application by coordinating with project team members to assure a positive outcome.

A Java-based desktop application that is a point of sale (POS) and distribution system that combines the latest features and technology for simple and easy sales transaction processing, inventory control, customer management, and purchasing control.

• Implemented the automated build through Git, Jenkins, and Install4j
• Enhanced the application's features as per received feedback from customers
• Integrated FBI - NICS APIs to submit electronic NICS for firearm dealers