Syed Obaidullah Hassan Chishti

Built an LLM agent achieving 94% SQL generation accuracy on medical databases

Led the LLM team in end-to-end development of intelligent agents

Streamlined model performance and deployment processes for scalability

Planned project pipelines and guided OCR team in model development and deployment for accurate data extraction

Managed data labeling operations to ensure production of high-quality annotated datasets

Built a clinical governance chatbot using RAG based on authentic medical guidelines

Fine-tuned bge-small and Mixedbread AI embeddings to reach 0.83 recall@5

Deployed sentiment analysis systems on Kubernetes for scalable usage

Developed a speech-to-text AI assistant using Whisper, Melo TTS, and LLaMA3

Achieved 94% accuracy in information extraction using YOLOv8 and PaddleOCR (91%)

Enhanced document classification with Mixedbread AI, achieving 0.94 F1 score

Built a CNN-based Tone Sentiment Model using MEL Spectrograms, achieving 84% accuracy for client sentiment insights

Designed a BERT-based text sentiment model with 92% accuracy for better decision-making

Applied KMeans clustering and an RNN-based model to separate IVR in speaker diarization

Led development of a real-time analytics dashboard using Flask and Apache Superset for business departments

Configured PostgreSQL and MongoDB on Linux and integrated with Superset for data visualization