Aug 2022 – Present · 3 yrs 9 mos

• AI Engineering: Production LLM Systems (2023–Present)
• Architected and built production LLM summarization pipeline serving 15,000+ agents daily. Accuracy
• went from 60% to 95%+
• Designed Judge LLM evaluation framework: automated model assessment with 7x faster iteration
• cycles
• Built modular AI platform enabling 3+ teams to self-serve and ship their own AI features independently
• Production systems: CS Co-Pilot (AI summarization), Intelligent Voice Assistant (call deflection),
• Automated Quality Evaluation (Judge LLM)
• Coordinated 5 external vendor engineering teams on AI initiatives through technical leadership and architecture
• Coached engineers from onboarding through tech lead development; established OKR framework for the team
• 50+ user shadowing sessions . Designed systems based on real agent workflows, not assumptions
• Infrastructure & Platform (2022–2023)
• Scaled Workflow Engine: integrated AWS OpenSearch, reduced RDS load, implemented access control for broader adoption
• Led CI/CD migration to Harness across entire track . Improved pipeline reliability for 90% of services
• Genius VIP Routing: architected priority routing for high-value customers across message and phone
• channels
• Technologies: Python, Java, TypeScript, LLM APIs (OpenAI, Claude, Gemini), LangChain, LangGraph, AWS (OpenSearch, RDS, Lambda), Docker, Kubernetes, Terraform, Grafana, Kibana

Jan 2020 – Aug 2022 · 2 yrs 7 mos

• Developed and documented multiple APIs, playing a pivotal role in the successful launch of a new customer support MVP. Demonstrated proficiency in full-stack development, showcasing expertise in Java, Perl, Protobuf, and Apigee.
• Collaborated seamlessly within the team, offering support and showcasing strong ownership and drive to deliver critical project components on time. Actively participated in code reviews, providing valuable feedback on merge requests to ensure code quality.
• Took initiative in setting up capacity testing and DQS access for Pega, displaying project management and problem-solving skills, ensuring robust performance and data quality.
• Efficiently collaborated with cross-functional teams, delivering high-quality results that highlighted excellent communication and teamwork abilities.
• Technologies Used: Java, Perl, JavaScript, Apigee, Locust, Docker, Kubernetes, Camunda, MySQL, Kibana, Grafana, AWS.

Dec 2016 – Nov 2019 · 2 yrs 11 mos · Bengaluru, Karnataka, India

• Helped scaling personalised education across ~1500 schools, ~500K students, ~20K teachers.
• ➢ Led the design & development of the core engine that personalized worksheets and question papers based on students' skills, enabling personalized education that catered to each student's individual needs and abilities.
• ➢ Utilized Domain Driven Design (DDD) to map out the entire domain, ensuring a collaborative approach to the development process that involved all stakeholders.
• ➢ Developed three applications: the Teacher App, School Admin App, and Parent App, which provided various functionalities to teachers, administrators, and parents, such as lesson plans, time tables, progress tracking, report generation, and communication.
• ➢ Led development of the Teacher App (Android), which facilitated teaching in the classroom, providing prompts to ask relevant questions, suggesting which student to ask, and allowing teachers to mark teaching as completed, send homework, take photos, and share them with the Parent App.
• ➢ Led development of the School Admin App (Web), which gave administrators visibility into how much progress teachers were making and allowed for report generation and bulk messaging to parents, as well as access to the School Feed.
• ➢ Led development of the Parent App (Android/iOS), which provided a feed of school activities, shared photos, homework assignments, and suggested material to view or learn after school hours, promoting greater parent engagement and involvement in their child's education.
• ➢ Demonstrated strong leadership skills, including effective communication, delegation, and task prioritization, to ensure the successful completion of projects and the delivery of high-quality solutions that met the needs end-users.
• ➢ Technologies used: Java 11, Dropwizard, Postgres, AWS, Terraform, Android, JavaScript, React.js

May 2016 – Dec 2016 · 7 mos · Remote

• Mobifly Media is a London based Ad-Tech Startup, one of the fastest growing mobile advertising company focused on performance marketing in key markets like North America, Europe & South East Asia.
• ➢ Redesigned the Ad-Mediation architecture for a scalable advertising solution. The old system, developed in PHP was not scalable, did not let to add more advertisers.
• ➢ Designed & Developed a generic Ad Mediation Engine with Python, Flask, Celery, RabbitMQ to handle large number of advertisers, designed to have a dedicated queue & workers for each advertiser to retreive campaigns at scheduled intervals, convert into a standard format & push to publishing channels(HasOffers).
• ➢ Implemented Fraud Reporting tools
• ➢ Developed an Admin dashboard to manage advertisers & affiliates.
• ➢ The implementation was Test Driven from scratch.
• ➢ Setup CI / CD to run tests & deploy on every push to github.
• Backend: Python, Flask, RabbitMQ, Celery, MongoDB, ElasticSearch, Redis, GridFS, AWS
• Frontend: HTML5, CSS3, JS

Aug 2015 – Apr 2016 · 8 mos · Bengaluru Area, India

• Dwellbird is a Technology platform that facilitates Landlords & Tenants to deal with each other directly throughout the rental lifecycle.
• ➢ Understood Residential Real Estate space, the problems faced by Landlords, Tenants & Brokers. Tried building a solution to make Home Renting super easy.
• ➢ Designed & Developed a Native Android app for Dwellbird having features to List Properties,
• Search Properties, Schedule Visit, Apply for Rent, Report Issues, Message Landlord / Tenant,
• Generate Rental Agreement, Perform Check-In Inspection. (http://play.google.com/store/apps/details?id=com.dwellbird)
• ➢ The product was Test Driven from scratch.
• ➢ Setup CI / CD to run tests & deploy on every push to bitbucket.
• Backend: Python, Django, ElasticSearch, MySQL, AWS
• Frontend: Android, HTML5, CSS3

May 2013 – Jul 2015 · 2 yrs 2 mos · Bengaluru, India

• Recruiterbox is a SaaS Software which allows to easily receive & manage the hiring and applicant process.
• ➢ Joined as 3rd employee, had been there till team size was 30+, mentored new joinee, helped hiring people, built / re-wrote most of the features of Recruiterbox.
• ➢ Built Interview Scheduling feature, let’s to sync 2-way with Google Calendar
• ➢ Built Advanced Search & Query Parser(Plasticparser) for Recruiterbox
• (https://github.com/Aplopio/plasticparser)
• ➢ Built Public API for Recruiterbox (http://developers.recruiterbox.com/)
• ➢ Automated Server Provising with Puppet
• ➢ Mentored team members
• ➢ Enforced Test Driven Development
• Backend: Python, Django, RabbitMQ, Celery, ElasticSearch, AWS
• Frontend: Backbone.js, React.js, Jade, HTML5, CSS3

Sep 2011 – Dec 2012 · 1 yr 3 mos · Bangalore

• Built virtual hardware for early software development in C/C++