Narayan Singh

B.Tech Computer Science and Engineering Student

Ghaziabad, Uttar Pradesh, IndiaSoftware Engineering and Artificial Intelligence/Machine Learning

2roles

58skills

1education

7credentials

About

Final-year B.Tech CSE student graduating in 2027 with two software engineering and machine learning internships, production backend and REST API experience, and hands-on work building AI/ML-powered features.

Experience

Data Scientist Intern

Sabudh Foundation

Jun 2025 – Dec 2025 · Mohali, Punjab (Remote)

Engineered production ML pipelines and automated data-ingestion workflows in Python, SQL, and REST APIs, cutting data-processing time by 35% across 50K+ monthly transactions. Built a multi-agent RAG system using LangChain, LangGraph, and Pinecone vector search; ran controlled retrieval experiments that reduced hallucinations by 40% and lifted answer accuracy by 25%. Led end-to-end EDA on 10M+ record datasets, surfacing insights that shaped 3 product decisions and improved user engagement by 18%; set up A/B testing frameworks achieving 92% precision and 89% recall.

ML Engineer Intern

Jyesta Corporate Entity

Jan 2025 – Mar 2025 · Bengaluru, Karnataka (Remote)

Shipped 2 production GenAI features in Python, TensorFlow, and LLM APIs, serving 10K+ daily active users. Developed RAG pipelines with embedding-based retrieval and vector search, improving accuracy by 32% while bringing query latency under 200ms. Designed ETL pipelines processing 500GB+ of structured and unstructured data weekly; refined prompts and hallucination detection, raising model reliability from 73% to 91%.

Education

Guru Gobind Singh Indraprastha University (GTBIT)

B.Tech, Computer Science and Engineering

June 2023 – June 2027

CGPA: 8.0 / 10.0. Relevant coursework: Data Structures & Algorithms, OOP, Operating Systems, Computer Networks, Database Systems, Software Engineering, Machine Learning, Deep Learning, Statistics.

Skills

PythonObject-Oriented ProgrammingData Structures and AlgorithmsJavaScriptTypeScriptSQLCC++BashFastAPIFlaskNode.jsRESTful API DesignMongoDBMySQLPostgreSQLLangChainLangGraphRetrieval-Augmented GenerationLLM APIsPrompt EngineeringAgentic WorkflowsScikit-learnTensorFlowKerasPyTorchPandasNumPyExploratory Data AnalysisFeature EngineeringA/B TestingETL PipelinesStatistical AnalysisDockerGitGitHubCI/CDMLflowDataiku MLOpsAWSAzureGCPJupyterVS CodeClaude CodeGitHub CopilotCursorPineconeFAISSChromaDBWeaviateVector SearchTensorFlow/Keras AutoencodersRandom ForestStreamlitTesseract OCRGroq WhisperFastAPI Development

Projects

AI-Powered Grievance Redressal System

Built a multimodal grievance platform for an Andhra Pradesh government hackathon supporting text, voice, image OCR, and documents; ranked in the top 10 of all submissions. Designed a LangGraph workflow classifying grievances into 11 categories, routing them across 13+ departments, assigning priority, and generating explainable decisions. Integrated Groq Whisper for Telugu/English voice transcription, Tesseract OCR, and RAG-powered resolution recommendations.

Multi-Agentic RAG System with AI Agents

Engineered a multi-agent architecture with a Brain routing agent that selects document retrieval, web search, or calculator tools based on user intent, reaching 94% routing accuracy. Optimized PDF parsing, Pinecone semantic search, BGE re-ranking, and answer synthesis to achieve NDCG@10 of 0.87 and a 78% improvement in answer relevance with source page citations. Integrated local and cloud LLMs with hallucination detection and deployed Streamlit, FastAPI, and CLI interfaces.

Credit Card Fraud Detection

Designed a hybrid fraud-detection pipeline using a TensorFlow/Keras autoencoder and Random Forest classifier, with separate training and inference pipelines, model persistence through Joblib, and an interactive Streamlit app for real-time predictions.