I'm Aditya Nandlal, an AI Agent Developer and multi-product engineer with approximately 10 years in the industry. I turn LLMs into useful systems with tools, MCP, RAG, approvals, automation, proprietary SaaS work, and production-grade software.
Five years of independent delivery. Five years building products at WovVTech.
Client ambiguity, product architecture, production constraints, and now agentic execution—the work has always been end to end.
Delivered 40+ private and personal projects across web products, admin systems, API integrations, automation, and business tooling. Owned discovery, architecture, implementation, deployment, and support.
Built multi-product systems across React, Next.js, Node.js, Python, data platforms, self-hosted infrastructure, and AI-assisted workflows, including admin and backend contributions to WovVMaps, a proprietary indoor-mapping and wayfinding product.
Publishes real, inspectable products spanning coding agents, MCP automation, rendering engines, backend platforms, secure self-hosted applications, and developer infrastructure.
Employer-owned product work cannot always be published on GitHub. The product is live and publicly verifiable; its implementation and source code remain confidential.
Contributed to the admin and backend systems used to build and operate interactive indoor maps: map-authoring workflows, floor/zone/POI data, route-ready venue models, and support for blue-dot, location-aware navigation experiences.
View live productVenue blueprints, floors, zones, points of interest, amenities, parking, searchable destinations, content, and branding.
Route-ready venue data for interactive navigation across malls, airports, hospitals, campuses, and other complex spaces.
Backend and data support for location-aware guidance, positioning flows, and live visitor navigation inside mapped venues.
Map experiences delivered through kiosks, web, mobile, QR and embedded integrations, with search and visitor analytics.
Contribution details are intentionally bounded. No private source, credentials, customer data, or internal architecture is disclosed.
Seven original public product builds from the GitHub profile—not tutorial clones, but opinionated tools with architecture, security, documentation, and delivery concerns.
A private-beta coding-agent foundation with a React workspace, Tauri/Rust safety layer, local Codex runtime, private LLM gateway, patch previews, and approval-gated actions. Some execution and installer paths remain in beta development.
A Python bridge and MCP server that lets coding agents control Blender through structured tools for scenes, assets, transforms, rendering, and export.
A low-level TypeScript 2D engine with explicit Canvas/WebGL pipelines and raw2d-mcp tooling for scene validation, generation, and agent automation.
A TypeScript headless CMS and API platform with multi-tenant isolation, workflow APIs, route permissions, automation tokens, and documented MCP contracts.
A self-hosted database access manager for PostgreSQL and MySQL with isolated adapters, restricted accounts, live catalog inspection, and a guarded SQL workspace.
White-label webmail built with strict TypeScript and provider adapters for JMAP and IMAP/SMTP, including protected setup, 2FA, hardened sessions, and Docker delivery.
An open-source Payload and Next.js server control plane for audited port operations over SSH.
The agent is only one layer. Reliable outcomes need contracts, state, permissions, observability, data, interfaces, and deployment.
AN / BESTMAA“Understand the system. Design the foundation. Ship the product. Verify the reality.”
Modular architecture, readable internals, practical automation, and products that can survive outside a demo.
Build · Play · Ship · RepeatLet's talk about agent workflows, MCP integrations, RAG systems, developer tools, or product infrastructure.