Rao Junaid

AWS AI/ML Engineer specializing in Generative and Agentic AI, with hands-on experience building production LLM systems on Amazon Bedrock, including multi-agent platforms, RAG solutions, intelligent document-processing pipelines, and natural-language analytics for healthcare, real estate, finance, and marketing clients. AWS Certified Generative AI Developer (Professional) with a strong foundation in classical ML, NLP, and computer vision.

AWS AI Engineer at Cloudelligent (2025-02 – Present)

• Built a production multi-agent real-estate intelligence platform (supervisor pattern, Strands Agents on Bedrock AgentCore) that queries Athena/Glue data lakes and live demographic APIs to generate investment-grade market reports across markets.
• Delivered enterprise GenAI on Amazon Bedrock (Claude/Titan/Nova): RAG chatbots on Knowledge Bases with OpenSearch k-NN vector stores, and an enterprise text-to-SQL solution with schema discovery and self-correcting queries (~91% accuracy).
• Engineered healthcare document-processing pipelines using Textract + Comprehend Medical + Bedrock to extract medical entities and auto-generate structured clinical SOAP notes from PDFs and transcripts.
• Built generative image-ad workflows with Amazon Nova Canvas and diffusion models, and conversational BI using Amazon QuickSight (SPICE, RLS, Q topics) and Amazon Q for multiple clients.
• Designed serverless, event-driven pipelines (Lambda, Step Functions, API Gateway); applied Bedrock Guardrails and CloudWatch observability; led LLM cost optimization (model routing, prompt caching) and mentored AI interns.

Machine Learning Engineer at DGB Technologies (2022-12 – 2023-10)

• Built supervised, unsupervised, and time-series models for demand forecasting and anomaly detection; trained and served models on SageMaker.
• Implemented Dockerized ML workflows (cut deployment time 40%) and Spark/SQL data pipelines; optimized models via feature engineering, PCA, and hyperparameter tuning.

Computer Vision Intern at Fenris Group (2021-07 – 2021-10)

• Annotated datasets, built OpenCV preprocessing and a CNN facial-recognition prototype; evaluated with IoU, precision, recall.

B.S. Software Engineering in AI specialization – University of Management and Technology, Lahore (2018 – 2022)