Portfolio

I recently worked on a structured SQL project using a comprehensive movie dataset to explore trends in film production, revenue, ratings, and genre distribution.

Key Objectives:

• Analyzed data related to movies, actors, genres, release years, and box office collections using MySQL.
• Built complex SQL queries involving joins, aggregations, subqueries, and window functions to answer business-driven questions such as:
  • What are the top-grossing genres over the last decade?
  • Which directors have the highest average ratings?
  • How do production budgets relate to box office revenue?

Outcomes:

• Discovered valuable patterns between budget, runtime, genre, and ratings.
• Provided insights that could help streaming platforms and studios optimize content planning.

Tools and Skills Used:
SQL, MySQL, Data Modeling, Joins, Aggregation, Subqueries, Window Functions, Data Cleaning

#SQLProject #MovieAnalytics #MySQL #DataAnalytics #SaiSingh #BusinessIntelligence #DataExploration #DataVisualization #GenAI

In this project, I focused on designing and testing prompt-driven solutions using OpenAI’s GPT models to solve real-world NLP tasks.

Project Objectives:

• Engineered effective zero-shot, few-shot, and instruction-based prompts for tasks such as summarization, sentiment analysis, classification, and text generation.
• Conducted prompt testing using structured input data and evaluated performance across multiple use cases.
• Integrated OpenAI’s API with Python workflows to automate response generation, fine-tune parameters, and measure accuracy.

Outcomes:

• Built a prompt library for rapid deployment of AI-powered solutions in business domains like customer service, content generation, and education.
• Improved task precision through prompt optimization and iterative testing.

Tools and Skills Used:
OpenAI API, Python, Prompt Engineering, NLP, LangChain, HuggingFace, JSON, API Automation

#PromptEngineering #OpenAI #NLP #GenAI #AIProjects #SaiSingh #DataAnalytics #PythonAI #LLMEngineering

I led a comprehensive analysis of the Superstore dataset to extract key insights on sales performance, customer behavior, and product trends. The project aimed to support business decisions with clear, data-driven recommendations.

Key Contributions:

• Cleaned and prepared the dataset using SQL and Python for accuracy and consistency.
• Performed detailed analysis of sales, profit, discount impact, and shipping efficiency across different regions, segments, and product categories.
• Developed interactive dashboards using Power BI, Tableau, and Zoho Analytics to visualize KPIs, trends, and performance comparisons.
• Applied correlation analysis and pivot modeling to uncover patterns and relationships influencing profitability.

Outcomes:

• Identified underperforming regions and product categories for targeted strategy revision.
• Delivered actionable insights that could help improve operational efficiency and revenue generation.

Tools and Skills Used:
SQL, Python, Excel, Power BI, Tableau, Zoho Analytics, Data Cleaning, Dashboard Design, Business Intelligence

#SuperstoreAnalysis #DataAnalytics #SQL #PowerBI #Tableau #ZohoAnalytics #BusinessIntelligence #SaiSingh #DataVisualization #GenAI

This project explored how Generative AI can enhance the traditional data analysis pipeline by automating insight generation, natural language reporting, and pattern detection.

Key Contributions:

• Integrated OpenAI’s GPT models into the data analytics workflow to generate natural language summaries, identify trends, and automate reporting tasks.
• Used Python and SQL for data cleaning, transformation, and querying across structured business datasets.
• Developed a prompt-based interface that allowed analysts to query data using plain English, powered by LangChain and OpenAI API.
• Compared AI-generated narratives with traditional BI dashboards built in Power BI and Zoho Analytics for validation.

Outcomes:

• Reduced time spent on manual analysis and report writing.
• Demonstrated how GenAI can improve accessibility and speed in business decision-making through human-like data interpretation.

Tools and Skills Used:
OpenAI API, LangChain, Python, SQL, Power BI, Zoho Analytics, Prompt Engineering, Data Visualization, NLP, GenAI Automation

#GenAI #DataAnalysis #PromptEngineering #OpenAI #LangChain #BusinessIntelligence #SaiSingh #AIinAnalytics #NLP #DataVisualization

This project focused on transforming raw data into clear, interactive, and actionable visual insights to support strategic decision-making.

Key Contributions:

• Designed dashboards and visual reports using Power BI, Tableau, and Zoho Analytics to present key metrics such as sales performance, customer segments, and product trends.
• Used SQL and Python to clean, filter, and transform datasets before visualization.
• Created custom visualizations (e.g., heatmaps, trend lines, regional maps) that enabled stakeholders to quickly interpret patterns and anomalies.
• Prioritized storytelling and user experience in dashboard design for non-technical audiences.

Outcomes:

• Delivered real-time, interactive dashboards that reduced reporting time and improved data-driven decision-making.
• Enabled business users to identify trends, monitor KPIs, and drill down into critical business areas with ease.

Tools and Skills Used:
Power BI, Tableau, Zoho Analytics, SQL, Python, Data Cleaning, Dashboard Design, Data Storytelling, Business Intelligence

#DataVisualization #DashboardDesign #PowerBI #Tableau #ZohoAnalytics #SQL #SaiSingh #BusinessIntelligence #Python #DataStorytelling

In this project, I developed a conversational AI chatbot capable of understanding and responding to user queries using transformer-based models from Hugging Face.

Key Contributions:

• Built a custom chatbot using Python and pretrained models from Hugging Face Transformers such as DialoGPT and BlenderBot.
• Integrated tokenization, attention masking, and inference logic to manage contextual conversations.
• Fine-tuned response flow using prompt formatting and model hyperparameter tuning for better coherence and relevance.
• Deployed the chatbot locally and tested it on use cases like customer support, FAQs, and general conversation.

Outcomes:

• Enabled human-like, multi-turn conversations with accurate context retention.
• Demonstrated how open-source LLMs can be adapted for domain-specific conversational agents.

Tools and Skills Used:
Python, Hugging Face Transformers, NLP, Dialog Systems, Tokenization, Model Inference, Prompt Design

#ChatbotDevelopment #HuggingFace #Python #NLP #ConversationalAI #SaiSingh #Transformers #GenAI #PromptEngineering