At Worldish, I build multilingual AI systems, including supervised fine-tuning pipelines for small language models and a healthcare conversational AI system for Tigrinya, Wolof, Bambara, and Arabic dialects. I also developed an Arabic dialect evaluation platform and applied reinforcement learning from human feedback to align model responses with human preferences.
At ThoughtSpot, I developed an incident response pipeline that routed Slack alerts and created Jira tickets, reducing mean time to resolution by approximately 60%. I also built Airflow failure detection and triage systems using anomaly detection, LangChain, and LangGraph, and contributed to fine-tuning Azure OpenAI models.
