Khawaja Ahmad

Senior ML Engineer

Zorvix Labs • Tiempo completo

Sep 2023 - Present • 2 yrs 8 mos

• Designed, developed, and deployed a RAG-based conversational AI system integrated with external messaging platforms (WhatsApp, Facebook Messenger), using LangChain and a vector database for context retrieval. • Developed a custom NLP chatbot using a sequence-to-sequence T5 model, which directly led to a 65% increase in measured customer satisfaction ratings. • Architected a multi-agent AI voice system using LangGraph and WebSockets, reducing token costs and context loss by 15% through task-specific agent orchestration. • Developed a RAG-based medical billing engine and automated end-to-end LLM workflows using n8n, Vector Databases, and Slack integrations. • Managed the fine-tuning and deployment of multiple large language models like Llama3-70b, Mistral onto resilient cloud infrastructure across AWS, GCP, and Azure. • Focused on model reliability by continuously monitoring performance in production and executing data-driven updates based on new input, including generating synthetic data to enhance image-based model accuracy. • Led technical knowledge-sharing and mentoring sessions to increase the overall AI proficiency of the engineering team. • Implemented Natural Language Processing (NLP) solutions to streamline and improve efficiency in the risk assessment workflow. • Achieved a 75% reduction in application crashes by leading the debugging, root-cause analysis, and issue resolution process across multiple platforms. • Collaborated effectively with non-technical stakeholders to refine and expand core enterprise information product datasets.

Machine Learning Engineer

Devnox AI • Tiempo completo

Sep 2020 - Sep 2023 • 3 yrs

• Conducted applied research focused on deep learning models for video surveillance and anomaly detection. • Played an active role in developing a cloud-based SaaS platform, utilizing Python's AWS SDK to manage and interact with various cloud infrastructure components. • Implemented RESTful APIs using Django and Flask to support document management, search functionality, and client-side data access. • Managed complex multimodal datasets and worked on 3D neural network architectures. • Successfully deployed optimized ML models directly onto resource-constrained Raspberry Pi devices for edge inference applications. • Built reliable CRUD APIs for a financial application, ensuring stability through automated testing frameworks.