Sumit Srivastava

• Focused research on a conversation personalization method called lexical alignment.
• Design and implement textual conversational agents that offer interactive and personalized explanations.
• Conducted two user research studies investigating the effects of lexical alignment in conversational agents on understandability, persuasiveness, trust, cognitive load, and information uptake.
• Currently developing a natural language generation engine integrated with a transformer-based Large Language Model (LLM) for controllable lexical alignment in conversational agents.
• Supervised five MSc and BSc students resulting in three papers and posters, and one journal paper is in progress.

• Developed the first MVP.
• Designed and developed the backend system of a self-serving programmatic advertisements Demand Side Platform (DSP), including third-party integrations (Xandr, Stripe, OpenAI), deployment (CI & CD), security (Amazon AWS VPC infrastructure), etc.
• Responsible for technology infrastructure policies.
• Responsible for ensuring security, and privacy of the technology infrastructure.

Main technologies: Java 8/17, Spring Boot, MySQL, AWS (Code pipeline, IAM, Cognito, Cloudwatch, VPC, Beanstalk, API Gateway, Lambda, Load balancer, Amplify), OpenAI (RAG, REST APIs), Google DialogFlow.

• Led and mentored a team of seven passionate software engineers.
• Led the development and maintenance of a cluster of seven software services.
• Overlooked the revamp of a real-time advertisements bidding service handling 1.5 million requests per
  second.
• - Migration to an event-driven architecture backed by an appropriate HTTP server.

- Achieved a latency of 2 ms and 1 ms for 99% and 95% of the requests, respectively.

- Migrated a component responsible for real-time lookups from files to PostgreSQL.

- Set up a low-latency and high throughput distributed system of PostgreSQL servers.

- I contributed to the improvement in performance of an open-source library to fetch data from PostgreSQL.

• Devised code and wiki documentation standards and overlooked its execution across the company.
  

Main technologies: Java, Amazon Web Services, Appnexus cloud, MySQL, PostgreSQL, Apache Camel, Java Mission Control, Perf-top, Docker, Vert.x, Google DV360

• Part of a team responsible for the development of a ‘Real Time Bid Modifier’ service handling up to 1.5 million
  requests per second.
• Revamped almost 70% of an interface service (back end and front end), updating the coding standards and upgrading to the latest libraries, and modularisation to improve efficiency and reduce redundancy.
• Designed and implemented a horizontally scalable service (Master-Slave) that compressed and transported logs up to 90 TB daily from the servers to AWS S3.

• Designed and implemented a service for fetching/locating data through different protocols, processing and storing them on AWS S3 and Redshift database.

- For 12+ different protocols.

- For 30+ data sources with different data types and custom processing.

• Mentored new contributors throughout my service at MiQ Digital after moving to other roles.
  Main technologies: Java, Apache Camel, AWS S3, AWS Redshift

• Worked on a project comprising a mobile advertising platform and central point currency system where users earn points to watch ads and redeem them for various rewards.
• Contributed to critical enhancements, bug fixes, and the addition of new features across different portals
  related to the project.
  

Main technologies: Ruby on Rails, Cassandra, MySQL, HTML, CSS, JavaScript.