Islāmābād, Pakistan
As a Senior AI and Machine Learning Engineer, I am currently working at RevelAI Health, where I design and deploy advanced AI-driven voice and chat agents for the healthcare sector. My work focuses on building reliable, efficient, and context-aware AI solutions that support care teams in improving patient interactions and operational workflows. I bring hands-on expertise in a modern AI tech stack, including Python, LangGraph, LangChain, Retell AI, HuggingFace, PyTorch, NVIDIA NeMo, Whisper, GPT-4o, Mistral, BERT, RoBERTa, RAG pipelines, Docker, AWS, GCP Vertex AI, and MongoDB. My experience covers the complete lifecycle of AI applications — from designing data pipelines and training models to optimizing deployment infrastructure and monitoring production systems. Collaborating across teams and geographies, I ensure solutions are delivered to a high standard and meet both technical and business needs. I adapt quickly to emerging technologies, approach challenges with a problem-solving mindset, and am motivated to contribute to impactful AI projects that push innovation forward.
- Designed and developed an agentic RAG-based chat assistant using LangGraph, LangChain, and Azure Openai & Bedrock Sonnet for healthcare applications. - Built voice agents with RetellAI to streamline inbound call handling and integrate call summaries into dashboards. - Created a MongoDB Atlas-based retrieval backend with versioned prompt tracking, evaluation and observability tools (Langfuse) to support debugging and traceability.
- Developed a speech-to-text and diarization pipeline using NVIDIA NeMo and Whisper for clinical encounter processing. - Enhanced RAG-based summaries and chat workflows using GPT-3.5, GPT-4o, and Mistral for medical information retrieval. - Implemented and scaled a Qdrant vector store for evidence-based biomedical literature search and integrated monitoring tools for model performance.
- Built and trained Classification and Regression models for medical insurance claims using BERT-based models - Trained customized models for Entity recognition and semantic clustering - Implemented Data augmentation using classical NLP techniques - Deployed finetuned pytorch-based ML models to cloud-based services (mainly GCP VertexAI)
- Built NLP pipelines using transformer models for entity extraction and classification in compliance automation. - Deployed AI models on AWS EC2, improving efficiency in processing regulatory documentation.