Krishna Narkhede
Data Science Intern at Samfocus Technologies Pvt. Ltd. Skills include Python, Java, C, SQL, Java Script, HTML.
Experience
Data Science Intern
Samfocus Technologies Pvt. Ltd
pune india
- Built LSTM-based Deep Learning models for financial forecasting using historical stock datasets and sequential analysis.
- Improved prediction accuracy using Feature Engineering techniques including SMA, RSI, MACD, and Min Max scaling.
- Optimized Model Tuning using Random Search, dropout optimization & evaluated through RMSE, MAE, & R2 metrics.
- Built analytics dashboards and visualization applications using Flask & Streamlit with automated workflow integration.
- Strengthened Backend Architecture and Scalability through a improved deployment workflows and testing pipelines.
Education
Vishwakarma Institute of Technology, Pune
Bachelor of Technology (B.Tech)
CPI: 8.64
Skills
PythonJavaCSQLJavaScriptHTMLCSSLLMNLPOCREDAData AnalyticsPandasNumPyStreamlitBootstrapMatplotlibOllamaMongoDBDBMSGitGitHubVS CodeExcelPower BIETL PipelinesData Visualization
ACHIEVEMENTS
Research Paper Presentation
STAI 2026
Presented a research paper on Optimized LSTM Networks for Robust Stock Price Forecasting at STAI 2026.
Best Pitch Award
Inspiron 4.0 National Hackathon
Secured the Best Pitch Award among 300+ competing teams at the prestigious Inspiron 4.0 National Hackathon.
IBM Certification
Data Analysis & Business Intelligence
Earned IBM Certification in Data Analysis & Business Intelligence fundamentals, focusing on analytics & visualization.
KEY PROJECTS
InsureQ-Insurance Document QA System
HackRx 6.0
Built a RAG pipeline using BGE embeddings and ChromaDB for semantic search and retrieval of document content. • Integrated Llama3 to generate accurate, context-aware responses from retrieved documents while reducing hallucinations. • Automated processing PDF, DOCX & TXT files through text extraction, chunking, metadata management, and indexing. • Developed a FastAPI backend with structured JSON outputs, confidence scoring, and intelligent query capabilities.
AP Invoice Automation System
Inspiron 4.0 National Hackathon
Designed invoice workflows leveraging AI-powered automation with OCR, document parsing and extraction pipelines. • Implemented Computer Vision techniques for invoice segmentation, structured field extraction & validation processes. • Built ETL pipelines to extract vendor data, taxation details & invoice identifiers via Automated Document processing. • Manual finance operations reduced by 80% through workflow optimization and intelligent automation frameworks. • Developed a scalable AP invoice platform integrating OCR, validation & extraction modules with end-to-end data pipelines.
AI Powered Data Processor
Developed a multilingual data analytics assistant powered by LLM-based query processing for MySQL & CSV datasets. • Enabled Natural Language-SQL conversion using Ollama Llama3 to automate database querying & analytical workflows. • Designed Streamlit dashboards featuring dynamic visualizations & automated chart generation for data-driven insights. • Integrated file processing, MySQL connectivity & multilingual support through Scalable Backend Analytics Pipeline.
Relevant Courses
Coursework
Machine Learning, Deep Learning, Data Science, OOPS using Java, Artificial Intelligence, Web Development
Leadership & Activities
Leadership & Activities
Mentored 30+ students for CII Eaton Changing Gears 7.0 leadership development initiative, fostering professional growth. • Participated in Technical competitions & collaborative engineering projects, applying analytical & problem-solving skills.