Humaira Tayyab

● Designed and delivered LLM-powered features for enterprise clients in retail, healthcare, and SaaS, covering document intelligence, conversational assistants, and content generation workflows. ● Built production RAG pipelines using LangChain, OpenAI APIs, and Pinecone, improving answer relevance on internal knowledge bases and reducing manual document lookup time by roughly 40%. ● Developed and fine-tuned ML models in PyTorch and scikit-learn for recommendation, classification, and demand forecasting use cases on large client datasets. ● Architected scalable inference services with FastAPI, Celery, and Redis, supporting high-volume API traffic with stable latency under load. ● Reduced model inference costs by approximately 30% through prompt optimization, response caching, batching, and right-sizing of model selection per use case. ● Implemented MLOps workflows with MLflow, Docker, Kubernetes, and GitHub Actions, cutting model deployment turnaround from days to hours. ● Built evaluation frameworks for LLM outputs covering accuracy, hallucination checks, and regression testing before each release. ● Integrated AI capabilities into existing client platforms through REST APIs, webhooks, and event-driven microservices on AWS. ● Developed React.js and TypeScript interfaces for AI tooling, including admin consoles, analytics dashboards, and human-in-the-loop review screens. ● Designed PostgreSQL and pgvector schemas for embeddings storage, semantic search, and audit logging of AI interactions. ● Implemented secure access patterns using OAuth 2.0, JWT, and role-based permissions across AI service endpoints. ● Partnered with product managers, data scientists, and client stakeholders to scope AI roadmaps and translate requirements into delivery milestones. ● Mentored mid-level engineers on LLM application patterns, code quality, and testing practices through reviews and pairing sessions.

● Developed full stack web applications for US-based clients across e-commerce, SaaS, marketing automation, and professional services using Python, Django, Flask, React.js, and PostgreSQL. ● Built and deployed machine learning features including product recommendation engines, churn prediction models, and user behavior analytics using scikit-learn and TensorFlow. ● Created REST APIs and backend services supporting catalog management, checkout flows, reporting dashboards, and CRM integrations. ● Improved API response times by 35% through query optimization, indexing, caching strategies, and asynchronous task processing. ● Implemented AI-driven personalization for an e-commerce platform, contributing to measurable gains in engagement and conversion through data-backed UI changes. ● Built ETL and data preprocessing pipelines with Pandas and Airflow to feed analytics and model training workflows. ● Containerized applications with Docker and supported AWS deployments using EC2, S3, RDS, and Lambda. ● Set up automated testing with PyTest and CI pipelines, reducing production defects across client releases. ● Collaborated in Agile sprints with distributed teams, handling estimation, client demos, and technical documentation.

● Delivered custom web applications and automation tools for small business clients in retail and e-commerce. ● Worked on an outdoor and sports gear e-commerce platform, supporting AI-assisted product recommendations, dynamic content rendering, and customer engagement analytics. ● Contributed to real-time monitoring and A/B testing setups that informed data-driven UI and UX improvements. ● Built responsive frontend modules with JavaScript and React.js and backend logic with Python and Flask. ● Managed full project lifecycles independently, from requirements gathering through deployment and client handoff.