Session: From Idea to Impact - Building GenAI solution
In the rapidly evolving landscape of generative AI, transitioning from a promising concept to a reliable production system requires a methodical approach. This talk will explore the key components of building a generative AI system for production, starting with clearly defining the use case and problem statement to ensure alignment with customer needs. We will delve into the iterative processes of experimentation and model development, emphasizing the importance of robust evaluation frameworks to validate performance, scalability, and safety. Additionally, the session will highlight strategies for incorporating continuous customer feedback to refine the system and drive successful real-world deployments. Attendees will gain insights into best practices for bridging research and production, ensuring that innovative generative AI solutions are both effective and resilient in dynamic environments.
Bio
Narmatha Balasundaram
Narmatha Bala is an Engineering Manager at Microsoft with over 15 years’ experience. She is an engineering leader with extensive background in various technologies in the financial services industry space. Narmatha loves the envisioning process with customer, ideating the ‘what could be possible’ and translating these into engineering solutions. Narmatha is passionate about bringing diverse perspectives and an active contributor, organizer, and participant in many organizations, from hackathons to Women in Tech and is a passionate mentor. Outside of work, talk to her about escape rooms, oil painting, throwing pottery on the wheel!
Kavitha Govindarajulu
Kavitha Govindarajulu is a Principal Software Engineer in Microsoft. She has over 17 years' experience in design, development, testing, and
implementation of large-scale cloud software systems and enterprise architecture.
Kavitha is passionate about D&I and empowering women in tech and mentoring.
Outside of work, Kavitha enjoys reading, yoga and building Legos with her kids.