Tech Stack · Tools & Languages

About Me

I'm Darshan Solabannavar — a CSE undergrad at Sapthagiri NPS University driven by data and applied AI. I clean real-world datasets, build dashboards stakeholders trust, and ship agentic AI systems with RAG and multi-LLM orchestration. From B2B datasets tracked across live Power BI dashboards to a fully agentic assistant deployed in production — I don't just analyze data, I engineer outcomes.

5+
Metrics Tracked
8.3
CGPA
3-Day
National Hackathon
#Best
NammaSuraksha 2025
How the Agentic AI Assistant thinks
INPUT user query
RETRIEVE RAG lookup
REASON Groq agent loop
ACT tool-use
OUTPUT grounded answer

Skills

01
Data Analysis & EDA
Cleaning, transforming, and exploring real-world B2B datasets with Pandas, NumPy and Excel. Feature engineering and statistical analysis at Bold Analytics.
02
Dashboards & Visualization
Power BI and Tableau dashboards tracking 5+ business metrics for stakeholders. Delivered consulting-style insights in Deloitte's Data Analytics simulation.
03
Machine Learning
Supervised learning — regression, classification, gradient descent, regularization — with scikit-learn. Built Linear Regression from scratch in NumPy.
04
Agentic AI & RAG
Multi-step autonomous agent loops on the Groq API with tool-use and retrieval-augmented generation. Multi-LLM orchestration with LangGraph state machines.
05
Languages & Tools
Python, Java, SQL. Git, GitHub, Docker, Jupyter Notebook, Google Colab, VS Code, Hugging Face.

Projects

01
Agentic AI · Python
Agentic AI Assistant
Fully agentic chatbot on the Groq API with real-time tool-use — calculator, Python execution — plus a RAG module for document-grounded answers. Multi-step autonomous agent loop, deployed live on Streamlit.
PythonStreamlitGroq APIRAG
View Project ↗
02
Open Source · TauricResearch
TradingAgents
Contribution to an open-source multi-agent LLM trading framework (arXiv:2412.20138) — LangGraph state machines orchestrating specialized agents across GPT, Claude, Gemini, Grok and DeepSeek, plus Docker deployment and persistent decision logging.
PythonLangGraphMulti-LLMDocker
View Project ↗
03
Machine Learning · From Scratch
House Price Prediction
Linear Regression model built from scratch in NumPy — feature scaling, cost function and gradient descent with no ML libraries. Validated across multiple feature sets to quantify what feature engineering actually buys you.
PythonNumPy

Certifications

Presidency University · Karnataka State Police
NammaSuraksha 2025 — Best Award Winner
Apr 2025
Stanford / DeepLearning.AI — Andrew Ng
Supervised Machine Learning
Oct 2025
Forage
Deloitte Data Analytics Job Simulation
Feb 2026

Let's Build

Open to Data Analytics or AI/ML internships, collaboration on meaningful projects, and conversations about building things that actually matter.

Get In Touch
solbannavardarshan@gmail.com