EX-Founder & Senior AI Lead
Designed and deployed a production-oriented conversational intelligence platform processing approximately 300 conversations per day.
- Designed the end-to-end production architecture using Clean Architecture and microservices.
- Built a cloud-native asynchronous ML inference platform for large audio workloads.
- Integrated speech-to-text, speaker diarization, NLP, behavioral intelligence and LLM models.
- Developed production REST APIs using Python and FastAPI.
- Built a custom Rust-based TGI server for Llama 3.3 70B.
- Designed scalable model-serving architectures for concurrent inference.
- Built RAG systems using Milvus, embeddings, semantic chunking and reranking.
- Implemented CI/CD and MLOps practices including model versioning and reproducible deployments.
- Implemented observability using MLflow, LangSmith, Prometheus and OpenTelemetry.
- Worked directly with business stakeholders to transform AI capabilities into product features.