# Adarsh Pandey — Software Engineer, AI & Full Stack Developer > Adarsh Pandey is a software engineer, AI specialist and freelance full stack developer from India — AI products, scalable web platforms, Chrome extensions, IoT and B.Tech-level engineering projects built with React, Next.js, Vue, Python, Java and Spring Boot. Website: https://www.beadarsh.in/ Email: info@beadarsh.in X: https://x.com/be_adarshh Instagram: https://instagram.com/beadarshh GitHub: https://github.com/beadarshh LinkedIn: https://www.linkedin.com/in/beadarshh/ ## Services (freelance & hire) - Freelance full stack web development — React, Next.js, Vue, Node.js, Java and Spring Boot - AI product development — LLM integration, RAG, chatbots and AI automation - Software engineering and system design — APIs, databases and scalable backends - Chrome extension and desktop app development - IoT and computer-vision projects, including engineering and B.Tech project guidance - Web performance, SEO and creative front-end animation ## Skills ### Technical - Full-stack web development — React, Next.js, Vue, Node.js, Java and Spring Boot - AI and LLM integration — Gemini, Groq / Llama, retrieval-augmented generation with pgvector - Computer vision and speech — OpenCV, YOLOv8, text-to-speech - API architecture and system design for scalable platforms - Databases — PostgreSQL, Supabase, MySQL and MongoDB - Creative development — GSAP, Three.js, WebGL and Framer Motion animations - Chrome extensions (Manifest V3) and Electron desktop apps - IoT — ESP microcontrollers and smart sensors - Automation pipelines and complex service integrations - Web performance, SEO and accessibility ### Non-technical - Team leadership and task delegation - Mentoring beginners in web development and data analysis - Product thinking and UI/UX design in Figma - Stakeholder communication and presentations - Project planning and delivery under deadlines ### Tech stack Next.js, React, Node.js, Spring Boot, Java, Python, C++, HTML5, CSS3, JavaScript, TypeScript, PHP, MySQL, PostgreSQL, Supabase, Docker, MongoDB, Github, Git, Postman, Tailwind, Gemini, ChatGPT, Claude, VS Code, IntelliJ, Jupyter, Figma ## Experience ### Software Developer Engineer (SDE) — Drone Federation India (DFI) Aug 2025 - Present Spearheading the development of robust and scalable digital platforms for airspace management and drone compliance operations. ### Mentor — Hiked.in Mar 2025 - Present Guiding beginners in web development, project building, and data analysis while fostering career readiness and technical confidence. ### Software Dev Intern & Team Lead — Bluestock Fintech Aug 2024 - Oct 2024 Led a team in developing scalable fintech solutions, overseeing task delegation, progress tracking, and mentoring. ### Software Development Intern — Centre for Development of Telematics (C-DOT) Jun 2024 - Aug 2024 Designed and implemented user interfaces for a secure Certificate Authority (CA) Portal, integrating backend services for certificate issuance. ## Projects ### EasyMail URL: https://www.beadarsh.in/projects/easymail 2026 · Email Platform · Developer Tools A self-hosted email platform on Resend — send, receive and track mail for all your projects, with a keyed contact API that lets any website send mail without exposing a mail key. EasyMail is a personal email platform I built on top of Resend. It gives every project one place to send, receive and track mail: a threaded inbox, compose and reply with rich text and attachments, and analytics for delivery, opens, clicks and bounces. Websites send through it with a single API call. In EasyMail I generate an API key for a project and choose which website domains may use it, which address it sends from and where messages are delivered. The site then posts its contact form to the EasyMail contact API with that key, and the message lands in my inbox with the visitor set as the reply-to address. This portfolio's own contact form works this way. It runs entirely on free tiers — Resend, Vercel and Supabase — with a guard that blocks sends before the daily or monthly limit is reached, and a permanent archive in Postgres because Resend only keeps mail for 30 days. **The problem:** Every small site needs a contact form, and wiring each one to a mail provider means scattering keys, quotas and inboxes across projects. One platform with scoped, revocable keys keeps all of it in one place. **Key features** - Contact API: one POST with an API key sends a message from any website - Keys are bound to a project, a From address, a recipient and the website domains allowed to use them; revoke any key instantly - Threaded inbox, sent mail and conversations, with replies that thread correctly in Gmail - Compose, reply and forward with rich text and attachments - Analytics for volume, delivery, bounce, open and click rates, per project - Free-tier guard that counts sent and received mail and blocks sends before the cap - Permanent archive of every message and attachment in Postgres and Supabase Storage - Private admin accounts with signed sessions and sandboxed rendering of incoming HTML **Tech stack:** Next.js, React, Tailwind CSS, Drizzle ORM, Supabase, Resend, Recharts, Tiptap **Links** - [GitHub](https://github.com/beadarshh/EasyMail) ### NIDAR URL: https://www.beadarsh.in/projects/nidar 2025 · National Challenge · Event Platform The official platform for NIDAR — the National Innovation Challenge for Drone Application & Research by MeitY — serving 3,500+ participants. NIDAR (National Innovation Challenge for Drone Application & Research) is a national-level drone innovation challenge hosted by the Ministry of Electronics and Information Technology (MeitY) under the SwaYaan initiative. The platform runs the whole challenge end-to-end: teams register and manage their members, move through multiple evaluation phases, book their slots, and receive certificates — while organisers run judging, logistics and live monitoring from a central admin control centre. It was built to stay fast and reliable under heavy, spiky traffic around deadlines and result announcements. **The problem:** A national challenge with thousands of student teams involves registrations, multi-round judging, scheduling, travel and accommodation. Running that over spreadsheets and email does not scale — participants and organisers need one reliable place for all of it. **Key features** - Team registration and member management for 3,500+ participants - Multi-phase evaluation with scoring workflows and live leaderboards - Slot booking and scheduling for evaluation rounds - Digital certificates generated for participants - Accommodation and event logistics, including digital invoicing - Admin control centre for real-time monitoring and operations - Handles heavy traffic, averaging 7k–8k users **Tech stack:** React, Tailwind CSS, Supabase, Framer Motion ### Orbit - Team Task OS URL: https://www.beadarsh.in/projects/orbit 2026 · Productivity · Collaboration A team-centric task OS with List, Kanban, Calendar and Sprint views, real-time sync and push notifications. Orbit is an internal task "operating system" built around teams rather than individuals. Work is organised into teams, projects, sprints and categories, and everyone works in a scoped space — My Tasks, a Team, or a Project. The same tasks can be seen as a List, a Kanban board, a Calendar or a Sprint, and every change syncs live across everyone looking at it. A triage panel catches tasks that still need an owner or a due date, so nothing slips through. **The problem:** Teams juggling separate tools for to-dos, boards and schedules lose track of who is doing what and when — and updates arrive late or not at all. **Key features** - List, Kanban, Calendar and Sprint views over the same tasks - Scoped workspaces: My Tasks, Team and Project - Real-time updates across everyone viewing a task - In-app and native push notifications - Rich-text task descriptions and drag-and-drop organisation - Triage panel for unassigned or undated tasks - Shareable links that keep the exact view and filters - Role-based access for admins, team admins and members **Tech stack:** Next.js, React, Supabase, Zustand, TanStack Query, Tailwind CSS ### PixUI URL: https://www.beadarsh.in/projects/pixui 2025 · UI Library · Creative Development A library of 40+ animated, copy-paste React components — WebGL backgrounds, text effects, 3D scenes, navbars and galleries — each with a live preview and code. PixUI is a showcase and library of creative UI components for web designers and developers. Every component has its own page with a live, interactive preview and the full source code, ready to copy into a React project. The collection covers the pieces that make a site feel alive: animated backgrounds, text effects, cursor effects, 3D objects, navigation menus, galleries and page transitions — plus full themed landing pages and character-driven "creatives". **The problem:** Polished motion and WebGL effects take a long time to build from scratch, and most inspiration sites show the result without the code. PixUI puts the working component and its source side by side. **What's Inside** - Backgrounds — Galaxy, Lightning, Grid, Squares and an audio spectrum - Text effects — Variable Proximity, Text Pressure, Shuffle, Scrambled Text, Scroll Reveal, Text Cursor - Animations — Laser Flow, Magnet Lines, Ghost Cursor, Image Trail, Pixel and Page Transitions, Folder, Cubes - 3D — Blob Background, Glass Torus and Floating Shapes - Navigation — Bubble Menu, Card Nav, Glass Surface Navbar and image-title navbars - Galleries — Circular Gallery, Dome Gallery, Stack and Scroll Stack - Themed pages — Stranger Things hero, Merry Christmas and a snowfall landing page **Key features** - 40+ components, each with a live interactive preview - Full source code on every component page, ready to copy - WebGL and 3D effects built with OGL and Three.js - Smooth scrolling and motion with GSAP, Framer Motion and Lenis - Creatives section with character-driven animated pieces - Per-page SEO and fast routing between components **Tech stack:** React, GSAP, Three.js (R3F), OGL / WebGL, Framer Motion, Tailwind CSS **Links** - [Live Site](https://pixui.beadarsh.in/) - [GitHub](https://github.com/beadarshh/pix-ui) ### FIELD URL: https://www.beadarsh.in/projects/field 2025 · IoT · Smart Agriculture An IoT-enabled smart farming solution integrating automated irrigation with bio-floc pisciculture. FIELD is an IoT-enabled smart farming system that combines automated irrigation with bio-floc pisciculture (fish farming), letting both run as one integrated, data-driven operation. Field sensors stream soil moisture, water pH and weather data to a web dashboard, where farmers can monitor conditions in real time and control irrigation pumps without being on site. **The problem:** Manual irrigation and water-quality checks are time-consuming and error-prone, wasting water and putting both crops and fish stock at risk. **Key features** - Real-time monitoring of moisture, pH, and weather - Automated pump status control - Web dashboard for efficient agricultural operations - Integrated irrigation and bio-floc pisciculture management **Tech stack:** React, Spring Boot, JavaScript, ESP (IoT), Smart Sensors ### Insight AI URL: https://www.beadarsh.in/projects/insight-ai 2025 · AI · Analytics Dashboard An intelligent platform providing data-driven insights and AI-powered analytics. Insight AI is an analytics portal that turns raw data into clear, AI-powered insights. It presents key metrics through an interactive dashboard so users can spot trends and act on them quickly. The interface is built as a modern, component-driven React app focused on readability of data and a fast, responsive experience. **The problem:** Raw data is hard to interpret at a glance; teams need a dashboard that surfaces what matters and explains it. **Key features** - Interactive analytics dashboard - AI-generated, data-driven insights - Responsive, component-driven UI **Tech stack:** React, TypeScript, Vite, Tailwind CSS, shadcn/ui **Links** - [GitHub](https://github.com/beadarshh/insight-ai-dashboard-portal) ### Alex AI Bot URL: https://www.beadarsh.in/projects/alex-ai-bot 2024 · AI · Computer Vision · Speech An AI-powered virtual assistant simulating realistic interview scenarios through multi-modal analysis. Alex is an open-source virtual interview assistant that runs a mock interview and gives real-time feedback on posture and voice — like a smart mirror for your body language and speech clarity. It combines computer vision on the webcam feed with speech analysis, then uses the Gemini API to merge those signals into personalised, constructive feedback that is also read aloud for accessibility. **The problem:** Interview preparation is stressful and honest feedback on body language and delivery is hard to get. Alex offers a private, judgement-free way to practise. **Key features** - Computer vision behavior analysis via YOLOv8 - Natural language understanding via Gemini API - Real-time voice interaction capabilities - Speech feedback on tone, clarity, pace and filler words - Session comparison to track improvement over time **Tech stack:** Python, OpenCV, YOLOv8, Gemini API, pyttsx3, Streamlit **Links** - [GitHub](https://github.com/beadarshh/Alex_Interview_Bot) ## Tools & open source ### React Content Protection URL: https://www.beadarsh.in/projects/react-content-protection 2026 · Open Source · npm Package · Tool / open source A configurable React package that protects web content by blocking copying, right-clicks, printing, shortcuts and developer tools. React Content Protection is an npm package for React apps that need to make their content harder to lift. Wrap your app (or just the sensitive parts) and it blocks text selection and copying, right-click menus, printing, common shortcuts like save and view-source, and attempts to open developer tools. Every protection can be switched on or off through a single config, and protection can be enabled conditionally — for example, letting admins copy while regular users cannot. It works with plain React and the Next.js App Router. **The problem:** Paid courses, exam portals and proprietary dashboards often need basic content protection, but wiring up dozens of browser event listeners in every project is repetitive and easy to get wrong. **Key features** - Blocks copy/paste and text selection - Disables the right-click context menu - Blanks the page on print attempts - Blocks shortcuts like Ctrl+S, Ctrl+P and Ctrl+U - Detects and blocks developer tools and debuggers - Detects automated browsers such as Selenium and Puppeteer - Per-option config, conditional protection and off-by-default in local development - Works with React 16.8+ and the Next.js App Router **Tech stack:** React, TypeScript, tsup, npm **Links** - [npm](https://www.npmjs.com/package/@beadarshh/react-content-protection) - [GitHub](https://github.com/beadarshh/react-content-protection) ### BlurSecrets URL: https://www.beadarsh.in/projects/blur-secrets 2026 · Chrome Extension · Privacy · Tool / open source A privacy-first Chrome extension that automatically blurs passwords, API keys, tokens and personal data so screen shares only capture blurred pixels. BlurSecrets keeps sensitive information off your screen shares and recordings. It continuously scans the page — including content loaded later by React or Angular apps — and blurs anything that looks secret, right in the page, so Zoom, Meet, Teams or OBS only ever capture blurred pixels. Beyond automatic detection, you can point at any element to blur it permanently on that page, add your own patterns, and click a blurred value to reveal it for three seconds before it re-blurs itself. **The problem:** One careless tab during a screen share can leak an API key, password or customer data. Remembering to hide everything manually before every call is unreliable. **Key features** - Auto-detects cloud API keys (AWS, Google Cloud, Stripe, GitHub, Slack and more) - Blurs personal data: card numbers, SSNs, emails and phone numbers - Blurs password fields and .env-style secrets - Custom regex rules for your own IDs and tokens - Inspector mode (Alt+Shift+X) to blur any element; it stays blurred on that page on future visits - Click-to-reveal for 3 seconds, then auto re-blur - Master on/off toggle and per-site whitelist **Tech stack:** JavaScript, Chrome Extensions (Manifest V3), CSS **Links** - [GitHub](https://github.com/beadarshh/Blur-Secrets-Extensions) ### DiagramGen URL: https://www.beadarsh.in/projects/diagramgen 2026 · AI · Developer Tools · Tool / open source An AI Mermaid diagram generator — write a prompt between // markers and get a flowchart, sequence diagram or DFD rendered live. DiagramGen is a split-screen editor for Mermaid diagrams. Type a plain-English prompt between // markers, press Enter, and an LLM writes the Mermaid code for you; the preview renders it instantly. You can keep refining a diagram by describing the change ("add a retry step after login"), edit the code by hand, and export the result as SVG or a high-resolution PNG. **The problem:** Mermaid is great for diagrams in docs, but remembering its syntax slows you down. Describing the diagram in words is faster. **Key features** - Prompt-to-diagram: write // your idea // and press Enter - Update an existing diagram with a plain-English instruction - Live, auto-rendering preview next to the editor - Supports flowcharts, sequence diagrams and data-flow diagrams - Export to SVG or 3× high-resolution PNG **Tech stack:** React, Vite, Mermaid, Groq (Llama 3.3) **Links** - [Live Site](https://mermaid.beadarsh.in/) - [GitHub](https://github.com/beadarshh/mermaid-diagram-generator) ### SEO Agent URL: https://www.beadarsh.in/projects/seo-agent 2026 · AI · Chrome Extension · Tool / open source An AI SEO analyst in a Chrome extension — audits any page, rewrites weak headings and meta tags, and learns from the fixes you accept. SEO Agent is a Chrome extension backed by a Python service that performs professional SEO audits on whatever page you are viewing. It reads the page content and structure — title, meta tags and H1–H4 headings — and returns concrete, ready-to-use rewrites rather than a generic checklist. Suggestions you accept can be fed back into the agent's long-term memory, so its advice gets more aligned with your own optimisation style over time. A history view tracks scores and progress per project. **The problem:** Most SEO tools flag problems but leave the rewriting to you, and they never learn what kind of fixes actually work for your sites. **Key features** - Deep page scan of content, title, meta tags and H1–H4 structure - One-click AI rewrites for headings and meta tags - "Teach AI" — store accepted findings in vector memory (RAG) - Accept & track suggestions per project - History and two-week performance reports **Tech stack:** Chrome Extension, Python, FastAPI, Groq (Llama 3.3), Supabase pgvector **Links** - [GitHub](https://github.com/beadarshh/seo-agent) ### Sticky Notes URL: https://www.beadarsh.in/projects/sticky-notes 2026 · Desktop App · Productivity · Tool / open source A context-aware desktop sticky notes app — notes belong to the app you wrote them in and appear only when that app is in front. Sticky Notes is a lightweight Windows app built with Electron. Each note is tied to the application it was created in: switch to Chrome and your Chrome notes appear, switch to VS Code and you see your VS Code notes. Notes float above their app, survive restarts, and come with a small but smart editor that continues lists and closes brackets for you. **The problem:** Ordinary sticky notes pile up on the desktop with no context. Notes about a project are most useful next to the app you use for it — and out of the way everywhere else. **Key features** - App-aware: notes show only when their app is in the foreground - Always-on-top above the owning app - Saved automatically; persists across restarts - Smart editor: auto-continued numbered and bullet lists, tab indents, auto-closed brackets - Five pastel colour themes per note - System tray icon and launch on Windows start-up - Portable .exe — no installer needed **Tech stack:** Electron, JavaScript, HTML, CSS, Node.js **Links** - [GitHub](https://github.com/beadarshh/sticky-notes) ### Browser Mascot URL: https://www.beadarsh.in/projects/browser-mascot 2026 · Chrome Extension · Fun · Tool / open source A draggable mascot that lives on every webpage — it follows your cursor with its eyes and reacts when you poke it. Browser Mascot is a Chrome extension that puts a small animated character on every page you visit. It turns to look wherever your cursor goes, and reacts when you click it — a blink, a heart, a sparkle, or dizziness if you poke it too fast. You can drag it anywhere (it remembers where), pick from built-in characters or bring your own sprite sheets — the extension includes a prompt guide for generating them with any AI image tool — and switch between six visual styles. **Key features** - Smooth eight-direction cursor tracking without jitter - Click reactions with squash-bounce, plus a dizzy state on rapid clicks - Draggable, with position remembered across pages - Built-in characters or custom AI-generated sprite sheets - Six visual styles: colour, ink, sketch, riso, paper and pixel - Isolated rendering, so it never clashes with page styles **Tech stack:** JavaScript, Chrome Extensions (Manifest V3), Shadow DOM **Links** - [GitHub](https://github.com/beadarshh/mascot-buddy) - [Based on page-mascot](https://github.com/nilbuild/page-mascot)