Real Capture
SmartSDLC Gradio UI
Real screenshot captured from SmartSDLC while running Gradio modules for AI-assisted requirement analysis and code support.
Gradio UIRequirement ClassificationAI Assistant
Portfolio Index
Charan Katkam
I ship AI tools with measurable impact
Full Stack Developer who measures every feature shipped. I build AI-powered tools that reduce QA effort by 60% and catch requirement risks before they become bugs.
60%
QA Effort Reduced
95%+
PDF Classification
100+
E-commerce SKUs
About Me
I'm a passionate Full Stack Developer with a strong focus on AI integration and modern web technologies. Currently pursuing my B.Tech in Computer Science Engineering (Class of 2026) at G. Pullaiah College of Engineering and Technology, I combine academic knowledge with hands-on development to deliver impactful solutions.
I got into AI tooling because I wanted to see how much developer effort could be automated — and SmartSDLC was my answer to that question. My expertise spans React, Next.js, Node.js, and AI/ML implementations using TensorFlow and OpenCV. I build tools that make developers faster, focusing on measurable outcomes like reducing QA effort and catching requirement risks early.
Currently Learning: LangChain, RAG pipelines, and advanced prompt engineering for production AI systems.
📍
Location
Hyderabad, India
💻
Background
AI + Full Stack
🗣️
Languages
English, Telugu, Hindi
Skills & Technologies
A comprehensive toolkit for building modern, intelligent web applications
Programming
PGJavaPythonJavaScriptTypeScript
Frontend
UIHTMLCSSJavaScriptReactNext.jsTailwind CSSTypeScript
AI / ML
AIPythonTensorFlowScikit-learnPandasNumPyOpenCVPrompt Engineering
Backend & Tools
BENode.jsExpress.jsMongoDBMySQLGitGitHubVS CodeJupyterREST APIs
Additional Expertise
Responsive Design
REST API Integration
Database Design
Git & Version Control
Agile Teamwork
Real Product Media
Replaced stock visuals with project-authentic screenshots and a working SmartSDLC demo clip.
Real Capture
Abhi Jewels Product Page
Live screenshot from the production product listing page, including filters, SKU cards, pricing, and category breakdown.
Live StoreProduct FiltersE-commerce
Real Capture
Air Cursor Gesture Tracking
In-action frame showing real MediaPipe hand landmarks and cursor mapping overlay used for touchless pointer control.
OpenCVMediaPipeCursor Mapping
SmartSDLC Short Demo Video
Open Live Demo VideoExperience
Practical experience delivering AI and full-stack products with measurable impact
AI SDLC Intern
Internship Project
Internship📅 2025
Developed SmartSDLC, an AI-powered Gradio application that automates critical SDLC phases using IBM Granite 3.3 AI model.
Key Achievements:
- Built AI-powered requirement classification module from PDF documents using IBM Granite 3.3, automatically categorizing requirements by type (functional, non-functional, constraints)
- Developed automated code generation system that produces Python code from natural language descriptions, accelerating development workflow
- Implemented intelligent bug fixing module that analyzes error messages, identifies root causes, and generates corrected code with explanations
- Created Pytest-based test case generation system that automatically produces unit tests for given code, improving code coverage and quality
- Designed code summarization feature that converts technical code into clear natural language explanations for documentation
- Integrated AI chatbot for real-time developer guidance on SDLC best practices and technical queries
Technologies Used:
PythonGradioIBM Granite AIHugging FaceGoogle ColabPyPDF2Transformers
Project Links:
Freelance Full-Stack Developer
AJ Abhi Jewels
Freelance📅 2025
Delivered a premium jewelry e-commerce platform for AJ Abhi Jewels with production-ready architecture, conversion-focused UX, and secure checkout flow.
Key Achievements:
- Built a modern Next.js 14 storefront with luxury branding, responsive layouts, and Framer Motion interactions for product discovery
- Implemented product search, advanced filtering, wishlist, and persistent cart flows for smooth repeat shopping sessions
- Integrated secure Razorpay payment flow and real-time cart/price calculations across the purchase journey
- Optimized performance using image optimization and lazy loading, targeting under-2-second load times on critical browse paths
- Shipped SEO-ready metadata, analytics hooks, and Vercel deployment setup to support discoverability and business operations
Technologies Used:
Next.js 14TypeScriptTailwind CSSFramer MotionZustandPrismaPostgreSQLRazorpayVercel
Project Links:
Featured Projects
Built projects with measurable outcomes, real media proof, and direct demo access.
SmartSDLC
Demo ReadyLive Demo VideoAJ Abhi Jewels
LiveOpen Live DemoAir Cursor Using Hand Gestures
Demo ReadyGitHubMusic Streaming Website
Demo ReadyGitHubSmartSDLC
Demo ReadyAI-powered SDLC assistant that automates critical SDLC phases, reducing manual QA effort by around 60% and classifying requirements with 95%+ accuracy.
Impact Metrics:
- Reduced manual QA effort by around 60% via automated test-case generation.
- Reached 95%+ PDF requirement classification accuracy in module validation runs.
- Delivered 2-4 second response windows for key module actions.
Technologies:
PythonGradioIBM Granite AIHugging FaceTransformersPyPDF2Pytest
AJ Abhi Jewels
LiveFull-stack e-commerce platform that handles 100+ jewelry products with sub-2 second load time and secure Razorpay checkout for end-to-end order flow.
Impact Metrics:
- Handles 100+ SKU catalog navigation with category and price filters.
- Maintains sub-2 second load times on critical product browsing paths.
- Supports complete user flow: discovery -> cart -> payment -> order follow-up.
Technologies:
Next.jsReactNode.jsExpressMongoDBRazorpayTypeScript
Air Cursor Using Hand Gestures
Demo ReadyComputer vision project for touchless cursor control with real-time tracking at 20+ FPS and under 50ms gesture response latency.
Impact Metrics:
- Real-time landmark tracking pipeline for gesture-to-cursor mapping.
- Supports drawing, erase, and interaction actions through tracked hand states.
- Built as an accessibility-first interaction experiment for hands-free control.
Technologies:
PythonOpenCVMediaPipeNumPyPyAutoGUI
Music Streaming Website
Demo ReadyMusic platform with lazy-loaded UI supporting 50+ tracks and under 1.5 second initial load time.
Impact Metrics:
- Search and playlist interaction across 50+ tracks.
- Custom player controls for smoother in-page listening flow.
- Structured for fast static deployment on GitHub Pages/Netlify.
Technologies:
HTMLCSSJavaScriptWeb Audio API
Explore pinned repos, demos, and updated READMEs on GitHub.
View GitHub ActivityInsights
Technical write-ups that show build depth, tradeoff thinking, and communication clarity.
2026-02-22 | 6 min read
How I Built SmartSDLC with IBM Granite AI
A practical breakdown of requirement classification, code generation, bug fixing, test creation, and summarization in one Gradio workflow.
Designed a 6-module workflow in Gradio for requirement analysis to developer-ready output.
Reached 95%+ requirement classification accuracy in PDF-driven validation runs.
2026-02-21 | 5 min read
What I Learned Building an E-commerce Platform from Scratch
Architecture and UX lessons from shipping a catalog-first e-commerce app with filtering, checkout, and payment integration.
Structured product browsing for 100+ SKUs with category and price filtering.
Maintained sub-2 second load time on critical product discovery paths.
2026-02-20 | 4 min read
From Webcam Gestures to Cursor Control: Air Cursor Notes
Lessons from real-time hand landmark detection, gesture mapping, smoothing, and usability tradeoffs in computer vision UX.
Implemented real-time hand landmark tracking with MediaPipe and OpenCV.
Mapped gesture states to cursor movement, drawing, and erase actions.
Let's Connect
Ready to collaborate on your next project? Let's discuss how we can build something amazing together.
Get In Touch
I'm always open to discussing new opportunities, interesting projects, or just having a chat about technology and innovation. Feel free to reach out!