🚀 Santhosh Guntupalli
Data Engineer | AI/ML Specialist | Cloud & BI Architect
🔗 LinkedIn | 📧 Email | 💻 GitHub |
🚀 Featured Projects
Explore some of my most innovative and technical work — covering AI, data pipelines, dashboards, and real-time systems.
⚡ GetGetLeads – AI Lead-to-booking & Digital Marketing Agent
**Ghostwriter Teams** is your local AI-powered creative department — built to help marketers, creators, and startups brainstorm, write, plan, and critique content like a full-scale agency.
Run locally with open-source models like LLaMA 2 or Mistral (via Ollama), it features five specialized agents working together to produce end-to-end campaign strategies, content calendars, rewrites, and branding insights — all without needing any paid API keys.
---
Ghostwriter Teams simulates a multi-agent marketing team where each AI agent specializes in a unique creative function:
- **Zara** – Creative Strategist
- **Max** – Content Architect
- **Mira** – Research Analyst
- **Eva** – Critic & Challenger
- **Leo** – Brand Guardian
You can brainstorm campaign ideas, generate structured content calendars, ask for rewrites, compare campaigns, and export your results — all through an intuitive Streamlit UI.
---
## ✨ Features
- **AI Agents as Roles** (Creative, Writer, Analyst, Critic, Brand Voice)
- **Full Content Calendar Generation** (Markdown tables with copy, dates, platforms)
- **Campaign Brainstorming + Rewrite Assistant**
- **Multilingual Output** (English, French, Hindi, Telugu, Japanese, and more)
- **Local LLM Runtime** with Ollama (LLaMA 2, Mistral, Gemma, etc.)
- **No API Keys Required**
- **Export Campaigns** as PDF, CSV, Markdown
- **Timeline Visualizations** with Altair & Plotly
- **Campaign Comparison & Dashboard View**
---
## 🚀 Example Use Case
> _"I want to launch an AI journaling app for moms. Help me plan 4 weeks of content on Twitter and LinkedIn."_
Ghostwriter responds with:
- A full **platform-specific calendar** of daily posts
- **Emotional hooks** and **CTA optimizations**
- **Eva’s critique** and **rewrite suggestions**
- **Mira’s insights** on trends, audience behavior, and timing
- **Leo’s brand voice check** to ensure tone alignment
Then, you export it all in Markdown, PDF, or CSV
**Tech Stack:** Streamlit · LangChain · Ollama · Local LLMs
---
## ✈️ GetGetPlaces – AI Travel Planning Agent
An intelligent NLP-powered travel planning platform with real-time weather, pricing, and smart routing.
Engineered a full-stack AI/ML solution for real-time personalized itinerary generation, synthesizing multi-source APIs (weather, hotel, transport) using time-series analysis, Location and optimization models, demonstrating scalable predictive analytics deployment
---
**Key Features:**
- Natural language trip planning: “Plan 5-day Miami trip with $1500”
- ARIMA price forecasting (flights, hotels)
- Route optimization + travel time estimates
- Smart daily itineraries with weather forecasts
- PostgreSQL backend for trip data
---
**Example Use:**
> “Plan a 7-day Orlando trip under $2500 focusing on theme parks and food” → Output: Day-wise schedule, attractions, maps, travel tips
---
**Stack:** PostgreSQL · ARIMA · OpenWeatherMap · LangChain
---
## 🎬 [Movie Recommendation System](https://mrs-sg-bfc2e6fa78db.herokuapp.com/)
Built a content-based movie recommendation system using Python, Scikit-Learn, TensorFlow, NLTK, Pandas, and NumPy, applying NLP and similarity matching algorithms to analyze structured/unstructured data and generate personalized predictions.
---
**Features:**
- Content-based filtering on movie descriptions
- Enter a movie → Get top 5 similar films instantly
- Web app interface with Heroku deployment
---
**Example Use:**
> Input: “The Matrix” → Output: Sci-fi action recommendations like “Equilibrium”, “Inception”, etc.
---
**Stack:** Flask · NLTK · Pandas · Scikit-learn · TF-IDF Vectorizer
---
## 🐍 Snake Game
A classic Snake game made using Python’s Tkinter GUI toolkit.
---
**Highlights:**
- Real-time movement and key-driven controls
- Score tracking and collision detection
- Minimalistic interface
---
**Example Use:**
> Launch the app → Arrow keys to control → Eat food to grow → Avoid crashing into walls
---
## 🏓 Pong Game AI – NEAT
Trains an AI to play Pong using the NEAT (NeuroEvolution of Augmenting Topologies) algorithm.
---
**Features:**
- Evolution of neural networks via reinforcement
- AI adapts its paddle response to opponent's movement
- Visualization of evolving agent skill
---
**Example Use:**
> Launch → Watch AI learn after several generations → Compete against the bot
---
**Stack:** Python · NEAT-Python
---
## 📈 Real-Time Stock Market Analysis System
Designed a live stock market trend prediction pipeline using AWS services (EC2, S3, Glue, Athena) and Kafka, building real-time data streaming architectures to deliver predictive investment insights.
---
**Features:**
- Live data ingestion via Kafka
- Transformation with AWS Glue and Spark
- Query using Athena and visualize in dashboards
- Use case: financial alerting, intraday trend monitoring
---
**Example Use:**
> Track tickers like AAPL, TSLA → Receive moving averages, trend shifts → Query on demand via Athena
---
**Stack:** AWS EC2 · Glue · Kafka · Athena · S3
---
🏠 [Back to Home](/Santhosh_Guntupalli.github.io/) | 📄 [Experience](/Santhosh_Guntupalli.github.io/experience.html) | 📫 [Contact](/Santhosh_Guntupalli.github.io/contact.html)