const vaibhav: Engineer = {
name: "Vaibhav Tulsian",
role: "Full Stack + AI/ML Engineer",
education: "B.Tech CSE (AI/ML) — PES University, Bengaluru",
location: "Bengaluru, India",
focus: ["Intelligent Product Engineering", "LLM Integration", "Scalable Web Systems"],
languages: ["TypeScript", "Python", "SQL"],
philosophy: "Ship products that think. Engineer systems that scale.",
};I am a Computer Science undergraduate specialising in Artificial Intelligence and Machine Learning at PES University, Bengaluru. My work sits at the intersection of full-stack product engineering and applied AI — I build end-to-end systems that are not only functional and scalable, but meaningfully intelligent.
My approach to engineering is product-first. I start from user problems, architect solutions with clean separation of concerns, and integrate AI where it creates genuine value — not novelty. I am particularly interested in LLM-powered applications, multimodal systems, and accessibility-driven technology that bridges digital gaps for underserved users.
I write primarily in TypeScript and Python, design systems with performance and maintainability in mind, and ship products that are production-ready from day one.
Open To →
Languages
Frontend
Backend & Databases
Cloud, DevOps & Tooling
| Domain | Proficiency | Details |
|---|---|---|
| Large Language Models | ████████░░ Advanced | Prompt engineering, RAG pipelines, LLM chaining, context management |
| Natural Language Processing | ███████░░░ Proficient | Text classification, summarisation, multilingual processing |
| AI-Powered Product Engineering | ████████░░ Advanced | Embedding LLMs into production-grade TypeScript/Python applications |
| Multimodal Systems | ██████░░░░ Intermediate | Audio synthesis, text-to-speech, prescription parsing |
| ML Model Integration | ███████░░░ Proficient | API-based model consumption, HuggingFace, OpenAI, Gemini SDKs |
| Accessibility AI | ███████░░░ Proficient | Building AI interfaces for elderly and low-literacy user segments |
◆ NovaLearn — Intelligent Learning & Teaching Platform
NovaLearn is a full-stack, AI-augmented educational platform that connects learners and educators through a cohesive, structured experience. Designed with a product mindset from the ground up, it delivers a seamless interface for course creation, content delivery, and progress tracking — built entirely in TypeScript for type-safe, maintainable code at scale.
| Attribute | Detail |
|---|---|
| Stack | TypeScript, Next.js, Node.js, PostgreSQL, Prisma ORM, Tailwind CSS |
| Scale | Multi-role platform — educators, learners, and administrators |
| Architecture | Monorepo, REST API, server-side rendering, role-based access control |
| Performance | Optimised query patterns via Prisma, SSR for fast initial loads |
| Security | JWT-based auth, input validation, protected API routes |
| Impact | Bridges the gap between content creators and structured learners |
| Repository | github.com/vaibhavtulsian/NovaLearn |
The platform was architected to eliminate the friction that typically exists between educators who want to share knowledge and learners who need structure. Rather than building a simple CRUD application, NovaLearn was designed as an extensible product — with a clean API boundary, normalised data models, and a UI that prioritises cognitive clarity over aesthetic novelty.
◆ Sanjeevani — AI-Powered Multilingual Medicine Companion
Sanjeevani is an accessibility-first AI platform designed to bridge the information gap between modern healthcare prescriptions and elderly, non-English-speaking, or low-literacy users. Users can query any medicine by name or upload a prescription and receive a clear, plain-language audio explanation in their preferred regional language — removing the dependency on a doctor or pharmacist for basic medication literacy.
| Attribute | Detail |
|---|---|
| Stack | TypeScript, Next.js, Node.js, LLM APIs (Gemini / OpenAI), Text-to-Speech APIs |
| Scale | Designed for pan-India deployment; multilingual support across 10+ languages |
| Architecture | AI pipeline — prescription parsing → LLM summarisation → TTS synthesis |
| Performance | Low-latency audio generation; optimised prompt design for accurate medical context |
| Security | No PII storage; ephemeral prescription processing; HTTPS-enforced |
| Impact | Enables elderly and rural users to safely understand their medication without assistance |
| Repository | github.com/vaibhavtulsian/Sanjeevani |
Sanjeevani addresses a critical and often overlooked problem — medication non-adherence and prescription misunderstanding among India's elderly population. The engineering challenge was not trivial: the system must correctly interpret medical terminology, simplify it without losing clinical accuracy, and then synthesise natural-sounding audio in languages like Hindi, Kannada, Tamil, and Telugu. This required careful prompt engineering, model selection, and a multi-stage AI pipeline built entirely in TypeScript.
| 🎓 B.Tech — Computer Science Engineering (AI/ML) | 2024 — 2028 |
PES University, Bengaluru
Pursuing a specialised undergraduate degree at one of India's premier engineering institutions, with a curriculum centered on machine learning theory, deep learning architectures, software engineering practice, and data systems. Simultaneously building production-grade projects that apply classroom concepts to real-world problems.
- Designing and shipping full-stack AI products as part of independent engineering practice
- Studying neural networks, NLP, computer vision, and statistical ML at the undergraduate level
- Contributing to open-source projects and maintaining a portfolio of production deployments
| Recognition | Details |
|---|---|
| 🛠 Production Deployments | Multiple full-stack applications deployed to production via Vercel with live user traffic |
| 🤖 AI Product Engineering | Built multilingual AI pipeline (Sanjeevani) addressing real-world healthcare accessibility |
| 📚 EdTech Platform | Architected and shipped NovaLearn — a complete learning platform from zero to production |
| 🎓 AI/ML Specialisation | Pursuing a dedicated AI/ML track at PES University — one of India's top engineering programs |
| 🌐 Full Stack Depth | End-to-end ownership across frontend, backend, database, and deployment layers |
# vaibhav.tulsian — current_state.yaml
learning:
- Advanced RAG architectures and vector database optimisation
- System design for distributed, AI-native applications
- Deep learning theory — transformers, attention mechanisms, fine-tuning
building:
- Expanding Sanjeevani with prescription image OCR and more regional languages
- New AI-powered product experiments at the healthcare and education intersection
- Open-source TypeScript utilities for LLM integration
exploring:
- Agentic AI systems and multi-agent orchestration frameworks
- Edge inference and on-device ML for accessibility applications
- MLOps pipelines and model serving at scale
open_to:
- Software Engineering Internships (Full Stack / AI / Backend)
- AI Engineering and Applied ML roles
- Open source collaboration on impactful TypeScript or Python projects
- Research partnerships in NLP, multimodal AI, or AI for social good
"The best engineers don't just write code — they engineer clarity from complexity."