Sai Krishna Rallabandi
Bio
Sai Krishna Rallabandi is Director, Data Science at Fidelity Investments. He earned his PhD at Carnegie Mellon's Language Technologies Institute where he contributed to the open-source FestVox and Flite speech-synthesis projects and Hear2Read, an Indian-language text-to-speech system for the visually impaired with 100,000+ downloads. He is also a Meta PhD Fellow with 60+ publications, he now builds Judith, a personal multi-agent that his family and friends use daily across WhatsApp, Telegram, and Discord.
Session (1)
Sai Krishna Rallabandi brings a rare combination of deep speech/NLP research (CMU PhD background) and applied financial AI at Fidelity Investments, with recent work on multi-agent LLM frameworks for generating persuasive investment reports and on preference-optimization (DPO) fine-tuning for conversational financial QA, making his session valuable for anyone building LLM systems for regulated or high-stakes domains.
GitHub
@saikrishnarallabandi- saikrishnarallabandi/Festival-Speech-Synthesis-System★ 3Modifications to Festival and Clustergen speech synthesis tools
- saikrishnarallabandi/Vocoding-Experiments★ 4Tracking vocoder experiments with the goal of revamping CMU's vocoder
- saikrishnarallabandi/jataayu★ 1LLM-backed AI agent security — inbound injection detection and outbound privacy protection
- saikrishnarallabandi/phdthesis★ 2Progress tracking for dissertation on de-entanglement for ubiquitous speech technologies
Recent writing (1)
Public activity researched automatically · as of Jun 2026