Conversation with Mohammad Omar (Founder of LXT) and @ianutile
summary
LXT describes building large training data pipelines for speech and conversational AI, serving major speech recognition firms, startups and researchers like SUPERB, with over 230 languages covered and a decade of deployments. Projects ramp from pilot to scale in days using crowds and managed pipelines that handle tens of thousands of contributors while enforcing secure QA. Claims focus on speed, accuracy and cost-effectiveness and include close co-working on pilots that produce deployable datasets.
AI-written summary · about
related talks
01h 02m[Full Workshop] Building Conversational AI Agents
10mBuilding Conversational AI Agents at Scale
21mTalking to others for you: Leveraging machines that speak the “human conversation” protocol
48mUnlocking LLM Potential: From Exploration to Integration
02mInterview with Mikhail Naumov
32mHarnessing Large Language Models for Enhanced Customer Experience
33mTranslate tech talks using open-source models and cloud-native Wasm
33mOpening Keynote: LLMs, between fantasy and reality - the poin...