Parveen Kashyap

Hybrid ANN-LSTM ensemble achieving 98.41% threat detection accuracy. Classified 15 attack types with 99.7% accuracy on 2.5M+ samples, integrating a self-healing remediation engine that decreased MTTR from 5 minutes to under 100ms via automated IP blocking.

Production RAG pipeline using LangChain LCEL and FAISS, achieving sub-second latency via token streaming and managing 10+ Gemini models with a resilient fallback to process 50-page docs in 5s.

Full-stack AI screening platform using TF-IDF and Cosine Similarity to rank resumes with 99% algorithmic accuracy. Engineered a secure SQLite backend with SHA256 password hashing and UUID salts.

• Engineered an AI governance system detecting 5 discrimination types across 100+ loan applications via a FastAPI/SQLite backend and Gemini Flash to enable real-time compliance auditing
• Developed a RAG medical chatbot achieving 95%+ accuracy by implementing Pinecone and Sentence Transformers for context-aware retrieval, transitioning early-stage prototypes into production-ready systems
• Built an optimized CV pipeline with 90%+ accuracy using YOLOv8 and EasyOCR, applying ROI-based filtering and inference tuning to automate data extraction from complex invoice tables

• Completed intensive, mentor-led programs involving 9 end-to-end ML projects across two organizations, focusing on model deployment and production-ready code architectures
• Developed an ASL recognition system achieving 98% test accuracy by implementing Transfer Learning (MobileNetV2) and custom CNNs in TensorFlow, deploying the model as a real-time Streamlit web application
• Optimized multi-class classification for forest cover types to 90.24% accuracy by benchmarking 7 algorithms using Scikit-Learn pipelines, GridSearchCV, and Extra Trees ensemble methods for model selection
• Engineered automated ML pipelines for fraud and medical datasets by implementing temporal feature engineering and Random Forest models, enabling batch processing and inference via Streamlit interfaces

• Specialized in Artificial Intelligence & Machine Learning
• CGPA: 7.76/10.0
• Focused on advanced ML algorithms, deep learning, computer vision, and AI systems development

• Percentage: 72.69%
• Built strong foundation in programming fundamentals, algorithms, data structures, and software development