Hemachandra Kolisetty

Software Engineer II

MICROSOFT•  May 2019 - August 2024•  Bengaluru, India

Developed an LLM-based generative AI assistant, GroupsGPT, tailored for M365 Groups platform with Azure OpenAI and Azure Cognitive Services to streamline group management through natural language prompts. Led a team of 4 engineers to redesign MS Teams’ Shared Channels’ REST APIs, which support over 2.5 million MAUs, to be future-proof, platform-agnostic, and scalable, enhancing interoperability and robustness. Architected a scalable backend service for roster management and CRUD operations for Loop workspace groups, supporting over 2.2M MAUs, using best practices for high-availability and efficient data handling. Revamped monolithic background assistants to microservice architecture, pioneering a POC for advanced scheduling pattern within Substrate, boosting scalability and efficiency for a platform with 250M MAUs and 22+ apps. Reduced video buffering time by 60% using ABR streaming with DASH and HLS, improving playback across diverse network conditions. Led research and backend optimization for streaming protocols and CDN configurations.

Software Engineer Intern

ACCENTURE•  May 2018 - July 2018•  Gurgaon, India

Engineered an efficient constraint programming solution for the capacitated vehicle routing problem with time windows and developed an end-to-end application that manages infeasible conditions and generates partial solutions.

Data Science Intern

iB HUBS•  May 2017 - July 2017•  Hyderabad, India

Developed a predictive meta-model that integrates Random Forests, Gradient Boosting, SVMs, and ANNs, leading to early detection and prevention of PID with 30% increase in accuracy and 20% reduction in false positives. Leveraged Association Rule Learning to discover hidden patterns in user behavior and medical data sets, enabling the development of a personalized, data-driven intervention system to identify high-risk patients early.