Open to opportunities

Sanjana Reddy Sangam

@sanjanareddysangam

Message

I build AI and backend systems that automate workflows and reduce latency.

India

Message

What I'm looking for

I'm looking for opportunities to build scalable backend and AI-driven products, work on automation and system design challenges, and grow through hands-on engineering collaboration.

I've built production-oriented systems at Amazon and Microsoft, from migrating core authentication APIs to reducing latency by 40% and creating an AI compliance-data pipeline for 36,000+ SaaS applications that reduced manual validation work by 60%.

I'm pursuing a B.Tech in Computer Science and Engineering at SRM University, where I've developed projects including a natural language-to-SQL generator with 95% accuracy and an automated PDF-to-Excel extraction tool. I also built a VGG16 skin-lesion classifier exceeding 90% validation accuracy and earned 2nd Runner-up at SRM University's CODE FEST.

Experience

Work history, roles, and key accomplishments

SDE Intern

Amazon

Jan 2026 - Jun 2026 (5 months)

Decommissioned a legacy orchestration service by migrating core APIs to the authentication service, reducing latency by 40% and migrating 5 clients using A/B testing frameworks. Authored HLD and AI-ready LLD specifications for device deregistration, and evaluated HTTP libraries to justify OkHttp migration, reducing login API latency by 30%.

Java Jersey OkHttp Dagger REST APIs A B Testing High Level Design Distributed Systems

SDE Intern

Microsoft

Jun 2025 - Aug 2025 (2 months)

Created an AI-driven system that extracted compliance data for 36,000+ SaaS applications, reducing manual validation work by 60% for 20+ analysts. Built a complete pipeline integrating Azure AI Agents, Bing Grounding, and Web Surfer for large-scale automation.

Azure AI Grounding Web Browsing AI Machine Learning Python Pipeline Development

Machine Learning Intern

NIT Trichy

Jun 2024 - Jul 2024 (1 month)

Developed a VGG16-based CNN using the HAM10000 dataset to classify skin lesions, achieving over 90% validation accuracy. Employed data augmentation, class balancing, and regularization to enhance model performance.

TensorFlow Keras Google Colab CNN Deep Learning Python