AI Animation

Parimal Kulkarni

AI & Data Science Enthusiast | GenAI Builder | LLM Specialist

View Projects Get In Touch

About Me

Hi, I'm Parimal Kulkarni, a driven and detail-oriented data science enthusiast currently pursuing my B.Tech in Computer Science at Symbiosis Institute of Technology. My passion lies at the intersection of Generative AI, NLP, and Machine Learning, with hands-on experience building robust, real-time systems that solve real-world problems.

Over the course of multiple internships at PowerCred (Singapore HQ), Britannia Industries, and Orinson Technologies, I've engineered full-stack GenAI applications, created scalable ML pipelines, and implemented LLM-powered assistants using LangChain, LLaMA3, FAISS, and RAG architectures.

Innovation

Building cutting-edge AI solutions that push boundaries

Efficiency

Optimizing processes and reducing manual effort

Precision

Developing solutions with high accuracy metrics

Scalability

Designing systems that handle real-world demands

Coding Visualization

Technical Skills

Programming Languages

Python, C/C++, SQL, MATLAB, JavaScript

Web Technologies

HTML, CSS, JavaScript, ReactJS, FastAPI, Streamlit

Data Science & ML

pandas, NumPy, scikit-learn, EDA, Feature Engineering, Statistical Analysis

Deep Learning & NLP

CNNs, RNNs, LSTMs, Transformers, PyTorch, TensorFlow, BERT

Generative AI

GPT-4, LLaMA 2/3, LangChain, RAG, FAISS, ChromaDB, Vector Databases

Tools & Deployment

AWS, GCP, Hugging Face, Docker, Git, PostgreSQL, SQLite

Featured Projects

94% Accurate

Advanced PDF Chat Assistant

Engineered a 94% accurate RAG pipeline with FAISS, boosting relevance by 40% using chunking, LLM chaining, and embedding models. Created web interface with real-time 100+ page PDF inference.

Python LangChain Groq API HuggingFace ReactJS
GitHub Demo
ArXiv W Wikipedia Web 25% Error ↓ Real-time

AI Research Assistant

Built GenAI agent with semantic routing, LLM selection, and real-time zero-shot search across ArXiv, Wikipedia, and Web. Enhanced answer grounding reducing LLM errors by 25%.

Python LangChain Groq API ArXiv Wikipedia
GitHub Demo
AI AI AI AI SQLite FAISS OAuth2

AgentForge

Comprehensive AI agent management system using FastAPI, SQLite, Groq API, and HuggingFace embeddings. Supports multi-source knowledge ingestion with FAISS vector storage.

FastAPI SQLite Groq API HuggingFace FAISS
GitHub Demo
<2s 1M+ Docs SOC 2 LLaMA3-70B

Google Drive AI Search

Built a GenAI Drive assistant using LangChain, Python, LLaMA3-70B, FAISS, and RAG for sub-2-second search across 1M+ documents with SOC 2-compliant architecture.

LangChain Python LLaMA3-70B FAISS RAG
GitHub Demo
Name Age Diag Med NLP 92% Accurate 1,500+ Records 70% Less Work Tesseract OpenCV

Medical Data Extraction OCR

Created OCR+NLP data pipeline for 1,500+ records/day at 92% accuracy, using DS/Algo, RegEx, and custom parsing logic. Cut manual effort by 70%.

Python Tesseract OCR OpenCV NLP FastAPI
GitHub Demo
SFR Rating Scale 1 9 Higher = More Stable ML Models 92% Acc RF DT SVM GB US 45% CN 20% UK 15% Python ML Mission Success Payload Analysis

Aerospace Investment Rating Analysis

Advanced ML-powered analysis of aerospace companies using SpaceFund Realty (SFR) ratings. Analyzes mission data, launch costs, payload capacity, and company stability to provide investment insights across the global space industry.

Python Pandas Scikit-learn Random Forest Decision Tree Matplotlib Seaborn GridSearchCV
GitHub Demo

Work Experience

Data Scientist - HITL Intern

PowerCred Technologies (Remote, Singapore HQ)

April 2025 - June 2025

  • Built a GenAI Drive assistant using LangChain, Python, LLaMA3-70B, FAISS, and RAG for sub-2-second search across 1M+ documents, improving precision by 35%
  • Engineered a SOC 2-compliant GCP architecture with OAuth 2.0, Cloud Functions, and metadata filters, serving 500+ daily queries
  • Developed AgentForge, a comprehensive AI agent management system with multi-source knowledge ingestion and conversational RAG pipelines

Data Science Intern

Britannia Industries Limited

February 2025 - March 2025

  • Developed a GenAI chatbot using GPT-4, LLaMA 2, ChromaDB, and custom domain embeddings, reducing average resolution time by 30%
  • Fine-tuned LLMs and optimized AI/ML workflows, improving accuracy by 25%
  • Performed sentiment analysis on 10K+ reviews using spaCy and NLTK, increasing engagement by 20%

Machine Learning Intern

Orinson Technologies Pvt Ltd

December 2024 - January 2025

  • Applied regression, classification, and clustering techniques to 200K+ records, enhancing model accuracy by 15%
  • Preprocessed data and fine-tuned ML models, reducing processing time by 20%

Get In Touch

Let's connect and collaborate on building the future with AI

parimalhkulkarni2@gmail.com +91 8087768244 Pune, India