Priyam Manna

Playwright, Pydantic, Python, FastAPI, and BeautifulSoup

• Created a FastAPI-based backend service for structured web scraping using Playwright and a static-first approach with a JavaScript rendering fallback.
• Interaction-aware scraping (scrolling, link clicks, and pagination tracking) was used to extract content from modern, JavaScript-heavy websites.
• Developed a dependable content parsing pipeline that divides DOM elements into semantic sections, applies noise filtering, and safely truncates raw HTML to prevent oversized responses.
• Clear contracts for metadata, sections, interactions, and error handling are made possible by well-defined Pydantic schemas for consistent API responses.

Fuzzy Matching, Python, Scikit-Learn, SpaCy, TF-IDF, Logistic Regression, Streamlit, and Joblib

• Developed a resume classification system that uses class-weighted logistic regression and TF-IDF features to predict over 100 job roles.
• Created a natural language processing skill extractor that uses regular expressions, SpaCy noun phrases, and cosine distance-based semantic similarity to identify over three hundred technical skills.
• Used a hybrid machine learning and rule-driven pipeline to map misspellings and variations, such as "reactjs" converted to "react," using fuzzy skill normalization.
• Developed a Streamlit web interface, a command-line tool, and a Python API for real-time resume analysis with lightweight model deployment using joblib.

Flask, NLP, BeautifulSoup, textblob, numpy

• Created a modular fake-news evaluation system combining NLP pipelines, sentiment analysis, and multi-factor credibility scoring.
• Developed real-time UI components with React and integrated WebSockets for real-time live score updates.
• Designed a distributed scraping and analysis pipeline using BeautifulSoup for batch URL processing and data retrieval.
• Deployed in Netlify using CI/CD, ensuring fast release cycles and zero-downtime updates.