Advay B.

Key Skills

Software

Other

Top Subject Matter

AI voice receptionist for inbound calls and appointment booking

Predictive modeling and analytics (ML workflow preparation and evaluation)

Healthcare assistant chatbot using RAG over medical PDFs

Top Data Types

Audio

Text

Document

Top Task Types

Transcription

Question Answering

Entity (NER) Classification

Freelancer Overview

RAG based Healthcare Assistant Chatbot (Project). Brings 2+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Other. Education includes Bachelor of Technology, SCMS School of Engineering and Technology (2025). AI-training focus includes data types such as Audio, Text, and Document and labeling workflows including Transcription, Evaluation, and Rating.

Labeling Experience

AI Solutions Developer — Klipto AI (March 2026 - Present)

OtherAudioTranscription

Built a real-time voice reception pipeline that converted inbound speech into text and then generated spoken responses. Implemented an end-to-end speech-to-text → LLM text processing → text-to-speech workflow for appointment booking automation. Optimized for low-latency conversational interactions suitable for real-time AI operation. • Speech-to-text transcription as the intermediate representation • LLM-driven text generation for dialogue and booking logic • Text-to-speech synthesis for end-user responses • Voice workflow orchestration and integration into an automation system

2026 - Present

AI and Data Science Internship — Techmaghi, Kochi (July 2025 - February 2026)

OtherText

Created and processed datasets for machine learning workflows using data cleaning and transformation steps. Trained and evaluated predictive models by applying feature engineering and hyperparameter tuning to improve performance metrics. Delivered outputs into analytics pipelines and stakeholder dashboards for decision support. • Data preprocessing using Pandas and NumPy for ML-ready datasets • Model evaluation and performance improvement via tuning • Integration of model outputs into analytics pipelines • Power BI dashboarding for communicating prediction metrics

2025 - 2026

AI Resume Screening Engine (Project)

OtherDocumentEntity (NER) Classification

Created an automated resume screening workflow that extracts candidate information and ranks candidates against job descriptions. Parsed resume documents and used NLP entity extraction plus weighted scoring based on TF-IDF similarity and relevance. Provided an interface for batch uploading, detailed result breakdowns, and configurable skill weighting. • Resume parsing for structured information retrieval • Entity extraction for identifying candidate skills and attributes • Weighted matching using TF-IDF cosine similarity and relevance scoring • Candidate ranking with CSV exports and configurable scoring

Not specified

Smart E-commerce Analytics (Project)

OtherText

Built an e-commerce analytics system that prepares and analyzes large datasets to support customer segmentation, churn prediction, and forecasting. Implemented model explainability and interactive visualizations to interpret predictive outcomes and communicate insights. Deployed the results as a live interactive web application for ongoing analysis and exploration. • Feature analysis and dataset preparation for prediction tasks • Churn model evaluation and interpretability using SHAP • Forecasting model training and performance reporting • Interactive Streamlit UI with parameter controls and exports

Not specified

RAG based Healthcare Assistant Chatbot (Project)

OtherDocumentQuestion Answering

Developed a retrieval-augmented generation (RAG) chatbot to answer medical questions grounded in a PDF knowledge base. Implemented vector search with FAISS and embedding generation using sentence-transformers to retrieve relevant passages. Added citation-style source attribution and conversational memory to reduce hallucinations and improve answer verifiability. • PDF knowledge-base grounding with retrieval • Source attribution using PDF page citations • MMR retrieval strategy to reduce irrelevant context • Conversational memory for multi-turn question answering

Not specified

Education

Work History