Manish Jain Email and Phone Number

Tanzu Insights (May 2022 - May 2024)- Worked on developing end-to-end OpenAI solution for generating Insights in Spring Java. Designed chunking tech to counter token limitations in GPT.- Conceived the algorithm and implemented complex way of representing Insights in Aggregated fashion in GraphQL call from Spring Boot — this API was the basis of product’s main UI.- Built Rule Engine with correlated rules, the first building block of Tanzu Insights.Designed and developed crucial GraphQL APIs and vital features; proficient in working with messaging queues, caches/hashes, databases, and container ecosystems.- Appreciated by colleagues and manager alike for creating some of the best documentation for new features, experiments, and reports.vRealize AI Cloud (Nov 2019 - Apr 2022)- Collaborated on building a complete inventory graph to manage vSphere devices utilized by vRAI, leveraging AWS Neptune and Gremlin.- Introduced and implemented reactive flow to the project to handle large numbers of concurrent connections with low latency-high throughput.- To ensure comprehensive testing of the API’s response with all resources, including databases and hashed data in place, initiated the first-ever series of TestContainers tests within the project.- Initiated & implemented E2E tests to check API response for varying inputs with TestNG & REST Assured.- Fulfilled my duties as an on-call PD engineer, serving as the go-to person for production issues and support.- Played an active role in designing and implementing the backend for RESTful APIs used in user interfaces as well as systems for queues and messaging to facilitate inter-service communication.vRealize Automation (Aug 2018 - Oct 2019)- Worked on Infrastructure as Code (IaC) while developing Cloud Assembly templates for vRealize Automation.- Solved bugs in Python for vRealize Automation.

Advisor: Dr. Hema A. MurthyMy master's thesis is on Implicit Systems for Spoken Language Identification. Implicit systems need no transcription or additional knowledge for processing given speech utterance, and thus these systems are easily scalable. The other goal of the thesis is to produce a robust technique that can accommodate new languages easily. The technique we developed is related to matrix factorization through which both scalability and robustness issues in language identification are solved.

Pod, da! was started in late 2015 as a podcast channel with a Facebook page. Pod, da! has been able to bring out the humor that exists there in the institute (IIT Madras) through observations in the form of text, image, or podcast. Every now and then, Pod, da! has given voice to problems faced by students and (especially) research scholars in the institute.

I worked on developing an end-to-end Spoken Language Identification system (DRDO) with IIT Madras' DON Lab headed by Dr. Hema Murthy. The work included:- Implemented and delivered UBM-GMM and SVD-SVM models in C, Python, and Bash. Used GNU Scientific Library for matrix operations in C.- Created a database from YouTube news recordings and processed it to retain language-only information.

- In the networks team for handling networking devices of SABIC Plastics. - Configured switches and solved tickets.