Full Stack Developer & AI/ML Expert

leetcode

gumroad

github

Akshat Tiwari  • Peerlist

With NextJS, TypeScript & TailwindCSS I have created a modern and good looking frontend for AI Agency,
This design has led to more client conversions and created an impact.

CherryAI - Automation in 5 clicks

Production-ready URL shortener with comprehensive analytics dashboard.

Key Features:
- Real-time analytics capturing click patterns, geographic data, device info
- JWT authentication with secure route protection
- Custom link creation and bulk operations
- QR code generation and CSV export functionality
- Redux Toolkit state management

Tech Stack: React 19, Node.js, Express, MongoDB, JWT, TailwindCSS, Redux Toolkit

Shawtly - URL Shortener with Analytics

Secure anonymous feedback platform with AI-powered conversation starters.

Key Features:
- NextAuth.js credential authentication with bcrypt encryption
- Google Gemini AI integration for real-time conversation generation
- Modern glassmorphism UI with dark mode support
- Real-time message management and toggle controls
- Unique profile links for each user

Impact: 40% improvement in user engagement

Tech Stack: Next.js 15, TypeScript, MongoDB, NextAuth.js, Gemini AI, TailwindCSS, Shadcn/ui, Framer Motion

Truly- Anonymous Feedback Platform

Full-stack AI SaaS platform integrating multiple state-of-the-art models (Gemini, ChatGPT, Mistral, Claude).

Key Features:
- Multi-model AI conversation hub with context retention
- AI image generation capabilities for creators and businesses
- Clerk authentication with end-to-end encryption
- Razorpay payment gateway integration
- Responsive React + Vite frontend with MongoDB backend

Tech Stack: React, Node.js, Express, MongoDB, Clerk Auth, Razorpay, AI APIs

AI Chat & Image Generation SaaS Platform

DhaanGo secured a Top 5 position at a national-level hackathon among 61 teams! at IIITM Gwalior Recognized for its groundbreaking potential.

🔗 Explore DhaanGo on GitHub : 
Delve into our technology stack and join our open-source initiative: https://github.com/aksh8t/DhaanGo.

📌 Project Overview  
DhaanGo empowers rural farmers by integrating technology into everyday agricultural practices. We aim to make farming more profitable, efficient, and sustainable through features that enhance accessibility and provide timely insights.

🔑 Key Features  
1. Dual Marketplace (Agriboost) – Enables direct farmer-to-business sales, product listings for agricultural supplies, and a streamlined e-commerce section.
2. MittiMitra - AI-Powered Chatbot – Delivers data-driven responses to farming questions, featuring voice integration and intelligent, agriculture-specific guidance.
3. Comprehensive Tools – Includes a profit calculator, MSP tracker, and weather forecasting for proactive financial and planning decisions.
4. Crop Management & Recommendations – Offers personalized fertilizer advice, crop disease identification, and yield optimization.
5. KhetWana - Community Platform – A peer-to-peer chat and resource-sharing hub for farmers to connect, support, and learn from each other.
6. Government Schemes & News Portal – Provides timely updates on government programs, eligibility, and agricultural news to help farmers stay informed.
7. SMS Alerts – Sends essential updates on MSP changes, pest outbreaks, and market prices, ensuring farmers receive critical information even offline.
8. API Integrations – Incorporates OpenWeather and News APIs for real-time weather, tomorrow.io, and more for a seamless experience.

💻 Tech Stack  
Frontend: HTML, CSS, JavaScript, Tailwind CSS  
Backend: PHP, Hack, SQL APIs, AJAX  
AI Integration: Python, Llama 3, Langchain, ElevenLabs   
Security: OAuth 2.0, PHP

DhaanGo : AI powered Web Platform to empower 

Excited to share my latest project: An Advanced Stock Price Prediction App! 📈💻

Key Features:
- 🧠 Multi-channel LSTM-based machine learning model and use of Deep learning with Keras & TensorFlow.
- 📊 Interactive visualizations of historical stock data and predictions with Matplotlib.
- 🔄 Real-time data fetching using yfinance
- 🌐 Streamlit-powered web interface for easy access to any Stock worldwide.
- Advanced feature engineering including moving averages, volatility metrics.

This model achieved impressive performance metrics:

Mean Squared Error (MSE): 0.0024
Root Mean Squared Error (RMSE): 0.0489
Mean Absolute Error (MAE): 0.0372
R-squared: 0.9876

These results demonstrate significant improvements over traditional single-channel models, with a 15% reduction in MAPE compared to our baseline.

Future Directions:
While this quantitative model has shown promising results, I'm excited about the next phase of development. My goal is to enhance the model's predictive power by incorporating non-quantifiable data through sentiment analysis, including:

News sentiment
Social media trends
Investor sentiment indicators

Github link: https://github.com/Aksh8t/Stock-Trend-Predictor-using-ML