NeuraApp — Interactive AI Learning Ecosystem
An AI-powered learning platform that helps developers and students master AI, ML, and software engineering through hands-on practice rather than passive video lectures.
I'm Devesh — an AI engineer and product builder working across AI, web and mobile. I enjoy taking ideas from early concepts to products people can actually use.

I'm an engineer focused on the intersection of technical rigor and real user utility. I don't look at software as merely lines of code or closed tickets in a backlog; I view it as an asset that should solve genuine problems and survive real-world scale.
My work spans the full spectrum of modern product engineering: from architecting LLM workflows and RAG pipelines with strict source verification, to building high-performance Next.js and FastAPI backends, to publishing responsive cross-platform mobile apps with Flutter.
This commitment to practical problem-solving led me to found NeuraApp. Watching students struggle with passive video lectures convinced me that tech education must be hands-on — featuring in-browser code execution sandboxes and intelligent, Socratic AI mentorship.
Each project represents hands-on engineering ownership — from problem formulation and system architecture to shipping production software users rely on.
An AI-powered learning platform that helps developers and students master AI, ML, and software engineering through hands-on practice rather than passive video lectures.
A cross-platform Flutter application for movie and TV discovery, trailers, and watchlist synchronization powered by the TMDB server API.
A Python utility that scans directories, classifies files by MIME types and creation metadata, and safely organizes cluttered filesystems with collision-proof timestamping.
Production-grade Retrieval-Augmented Generation architecture designed for zero-hallucination querying over multi-format technical documentation.
A disciplined engineering foundation developed through production builds. No superficial logo walls — only disciplines and tools actively applied to build software that lasts.
Most technical education forces students into passive video consumption. When they finally write code, they hit environment setup hurdles, receive cryptic syntax errors, and struggle without instant feedback. NeuraApp is built around active, guided execution.
NeuraApp transforms tech education into a personalized, hands-on experience. Learners write real code directly in sandboxed browser environments, receive instant Abstract Syntax Tree (AST) test verification, and learn with 'Ask Nea' — a lesson-aware Socratic AI mentor that explains concepts and nudges students forward without spoiling answers.
Zero Local Setup
In-browser sandboxes for immediate coding
AST Verification
Automated syntax & logic test suites
Ask Nea Mentor
Context-aware Socratic feedback
Core AI to LLMs
Structured engineering learning spines
Isolating untrusted user code execution in secure, low-latency container environments while delivering sub-second execution feedback.
Prompt-engineering and grounding the 'Ask Nea' assistant with lesson context, test-case AST rules, and strict guardrails so it teaches principles instead of simply emitting copy-paste solutions.
Structuring a progressive learning graph starting from Python and linear algebra up through PyTorch, Deep Learning, RAG, and production LLM orchestration.
A chronological record of verified engineering contributions, technical leadership, and open-source public work.
Conceived and engineered an AI-powered interactive learning platform helping students and developers master AI, ML, and software engineering.
Partnering with teams and building independent technology products spanning AI workflows, modern web platforms, and automated backend infrastructure.
Engineered responsive, cross-platform mobile applications for Android and iOS using Flutter and Dart.
Reflections from building NeuraApp, designing verifiable AI workflows, and scaling production software.
How we architected 'Ask Nea' to provide contextual debugging hints and conceptual explanations using AST validation and strict retrieval guardrails — without spoiling solutions.
The difference between a demo and a business asset — and how to scope the first release so founders learn fast without burning runway.
Onboarding, performance, and release discipline — practical choices that turn downloads into daily users and paying customers.
I'm always interested in thoughtful conversations, challenging engineering problems and meaningful collaborations.
Tell me what you are building, the problem being solved, or the technical collaboration you have in mind.