Yash Maheshwari

Built cloud IDE backend with FastAPI + PostgreSQL, async architecture, WebSocket streaming, and distributed workflow execution with checkpointing and fallback paths.

Implemented

GitHub API integration for automated codebase crawling and traversal — repository ingestion, branch-level indexing, and context retrieval pipelines across entire codebases. Engineered full backend data pipeline — chunking, vector indexing, reranking — for semantic code search; improved backend throughput by 30% via optimized async design.

Software Engineer Intern - Orydle - Remote

(2026-01 - 2026-04)

• Built cloud IDE backend with FastAPI + PostgreSQL, async architecture, WebSocket streaming, and distributed workflow execution with checkpointing and fallback paths.
• Implemented GitHub API integration for automated codebase crawling and traversal — repository ingestion, branch-level indexing, and context retrieval pipelines across entire codebases.
• Engineered full backend data pipeline — chunking, vector indexing, reranking — for semantic code search; improved backend throughput by 30% via optimized async design.
• Designed structured output schemas and prompt optimization strategies reducing external API cost by 25%; built fallback and retry logic for agent reliability.
• Architected microservices-based system with modular autonomous agent components, distributed task queues, and real-time streaming interfaces.

AI/ML Intern - DevX - Surat

(2025-12 - 2026-01)

• Built production ML forecasting engine (SARIMAX + Scikit-learn) with end-to-end feature engineering pipeline; improved prediction accuracy by 70%.
• Developed scalable FastAPI inference pipelines with async endpoints; implemented Pandas + NumPy preprocessing reducing data preparation time by 40%.
• Designed and maintained eval sets and regression tests to catch model drift pre-deployment; established structured quality scoring per workflow.
• Built automated business report generation system using Hugging Face Transformers; documented experiments, prompt templates, and evaluation results.
• Optimized PostgreSQL query patterns and data access layers for inference pipeline performance.

B.Tech - Computer Science and Engineering (Data Science) - Manipal University Jaipur (2027) Jaipur, Rajasthan