Shahid Hussain

Key Skills

Software

Google Cloud Vertex AI

Top Subject Matter

Conversational AI/NLP

Document Search/Retrieval Augmented Generation

Image-based Recommendation Systems

Top Data Types

Text

Image

Document

Top Task Types

Question Answering

Classification

Entity Ner Classification

Transcription

Freelancer Overview

Intent Classification Labeling – AI Autonomous Chatbot Platform. Brings 3+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal, Proprietary Tooling, and Google Cloud Vertex AI. AI-training focus includes data types such as Text, Document, and Image and labeling workflows including Classification, Entity (NER) Classification, and Transcription.

Labeling Experience

Image Classification Labeling for Hair Care Product Recommender

ImageClassification

Engineered a complete data labeling and inference pipeline for product images used in an AI recommender system. Managed feature extraction, classification, and prediction accuracy through labeled image datasets and barcode integration. Responsible for end-to-end label generation, confidence scoring, and consistency checks to achieve high system performance. • Integrated OpenAI API and Django for structured product classification labeling. • Oversaw labeling across a dataset of 500K+ products, ensuring 98% compatibility labeling. • Implemented confidence scoring mechanisms for label evaluation and improvement. • Delivered a scalable pipeline within a 6-week timeline with no external oversight.

Engineered a complete data labeling and inference pipeline for product images used in an AI recommender system. Managed feature extraction, classification, and prediction accuracy through labeled image datasets and barcode integration. Responsible for end-to-end label generation, confidence scoring, and consistency checks to achieve high system performance. • Integrated OpenAI API and Django for structured product classification labeling. • Oversaw labeling across a dataset of 500K+ products, ensuring 98% compatibility labeling. • Implemented confidence scoring mechanisms for label evaluation and improvement. • Delivered a scalable pipeline within a 6-week timeline with no external oversight.

2025 - 2025

RAG Chatbot Data Preparation and Evaluation (Custom Document RAG Chatbot)

DocumentQuestion AnsweringEntity Ner Classification

Developed a production Retrieval-Augmented Generation (RAG) chatbot utilizing LangChain and OpenAI over document corpora. Conducted data preparation including chunking, embedding selection, and vector index tuning to improve answer relevance. Evaluated and iterated on output quality via semantic search and reranking for optimized information retrieval. • Performed data labeling and preparation by segmenting Word/PDF documents for semantic search. • Engineered dense embedding pipelines and selected retrieval strategies for robust question answering. • Evaluated chatbot responses to measure and annotate answer relevance on internal benchmarks. • Utilized Weaviate and FastAPI to manage embedding, indexing, and conversational pipeline workflows.

Developed a production Retrieval-Augmented Generation (RAG) chatbot utilizing LangChain and OpenAI over document corpora. Conducted data preparation including chunking, embedding selection, and vector index tuning to improve answer relevance. Evaluated and iterated on output quality via semantic search and reranking for optimized information retrieval. • Performed data labeling and preparation by segmenting Word/PDF documents for semantic search. • Engineered dense embedding pipelines and selected retrieval strategies for robust question answering. • Evaluated chatbot responses to measure and annotate answer relevance on internal benchmarks. • Utilized Weaviate and FastAPI to manage embedding, indexing, and conversational pipeline workflows.

2025 - 2025

Audio Transcription Labeling for Quran Learning Application

Google Cloud Vertex AIAudioTranscription

Built a production-grade AI platform for audio pronunciation assessment using real-time speech-to-text labeling. Managed sync between audio inputs and transcribed outputs to enable accurate learning evaluation. Optimized labeling accuracy to 81.9% and reduced error rates by orchestrating automated transcription and label verification frameworks. • Applied Google Cloud Speech-to-Text for large-scale audio transcription. • Designed WebSocket-based annotation workflow for real-time labeling feedback. • Maintained robust labeling structures for 1,000+ simultaneous learners. • Improved sync and annotation precision, achieving measurable educational impact.

Built a production-grade AI platform for audio pronunciation assessment using real-time speech-to-text labeling. Managed sync between audio inputs and transcribed outputs to enable accurate learning evaluation. Optimized labeling accuracy to 81.9% and reduced error rates by orchestrating automated transcription and label verification frameworks. • Applied Google Cloud Speech-to-Text for large-scale audio transcription. • Designed WebSocket-based annotation workflow for real-time labeling feedback. • Maintained robust labeling structures for 1,000+ simultaneous learners. • Improved sync and annotation precision, achieving measurable educational impact.

2024 - 2025

Education

Work History