Ayush Mishra
Product Engineer focused on AI-native platforms, knowledge systems, and scalable SaaS architecture
Computational architect and product engineer building AI-native platforms for the AEC industry. My work sits at the intersection of architecture, parametric design, and software engineering, where I develop systems that turn complex building data, construction documents, and infrastructure knowledge into intelligent digital tools. With a background in architectural design and distributed systems, I focus on creating software that improves how the built environment is designed, analyzed, and managed.
Work Experience
- Contributed to the design and development of the Delhi Metro Skywalk (Ajmeri Gate) connecting New Delhi Railway Station with Delhi Metro transit infrastructure.
- Designing and creating React components for the application dashboard.
- Developed computational design and parametric workflows for complex infrastructure geometries and architectural systems.
- Built internal AEC software platforms and digital infrastructure to improve knowledge management, documentation intelligence, and design automation across architecture and construction workflows.
- Designed automation pipelines and computational tooling integrating architectural data, contracts, and project documentation.
Key Projects
Enterprise-grade AI document management system for AEC contractors. Processes hundreds of documents with thousands of pages using multi-provider AI (Gemini, GPT-4, Claude) with RAG architecture, pgvector semantic search, and multi-tenant RBAC. Serving 500+ users at NBCC, L&T with 80% reduction in document search time.
Social media platform for AEC professionals with revolutionary attribution graph connecting projects, contributors, and products. Features interactive product hotspots, ARQ AI visualization engine, ML-powered recommendations, real-time feed, jobs board, and comprehensive analytics. 2,500+ users, 2.3M monthly page views, 85% DAU.
8-stage 'Technical Moat' rendering pipeline orchestrator dynamically routing generative requests across Gemini 2.5 Flash/Pro, Google Veo, and Hunyuan3D. Stateful node-based workflow editor with RAG-style 'Pipeline Memory' using Gemini Vision. Ultra-low costs (~$0.04/render).
Education
M.Tech
B.C.A (D.Sys.)
B.Arch (Design)
Skills & Tools
Programming
Web & API
AI & Data
Spatial & 3D
Databases & Cloud
Languages
References
Selected work and references available upon request from Studio ISA and Inheritage Foundation leadership.
