Available for Projects

Hi, I'm Shourya.

Building AI & ML Systems.

23-year-old Data Scientist & AI Developer. Specializing in GenAI, Agentic AI, Machine Learning, and Data Science — architecting intelligent multi-agent systems, RAG pipelines, and LLM-powered applications.

print("Welcome to my portfolio!")

visibility View My Work download Download Resume mail Get in Touch

180M+

Records Processed

12M+

Documents Analyzed

92.5%

Model Accuracy

88%

Cost Reduction

Technologies I Work With

terminalPython databaseSQL neurologyPyTorch psychologyLangChain cloudAWS deployed_codeDocker hubFastAPI smart_toyRAG

Technical Arsenal

code

Programming & Databases

Python, SQL

neurology

Data Science & Libraries

Pandas, NumPy, Scikit-learn, PyTorch, Keras, matplotlib, OpenCV

bar_chart

ML Techniques

NLP, ETL, Statistical Analysis, EDA, Regression, Classification, Clustering

psychology

Generative AI

LLMs, LangChain, RAG, Agentic Workflows, Langgraph

cloud

Cloud & MLOps

AWS (SageMaker, S3, Bedrock, EC2, ECR), Docker, FastAPI, Streamlit, Mlflow

build

Tools

Git, GitHub, Power BI, Advanced Excel, HTML, CSS, Data Storytelling, Web Scraping

Professional Journey

May 2025 - Present

KOTAK LIFE

Data Scientist

KOTAK LIFE | May 2025 - Present

Developed and deployed AI-driven underwriting and decision systems on AWS (SageMaker, S3, EC2), using Langgraph enabling scalable experimentation, training, and inference.
Designed a multi-agent RAG evaluation pipeline using Agno, HuggingFace embeddings, LanceDB, LLM reasoning, improving accuracy of insurance pitch assessments and recommendations.
Engineered a large-scale data processing pipeline handling 180M+ insurance records and 100+ Excel QRFs, performing extraction, deduplication, clustering, and enrichment using Python, Pandas, KMeans, hashing, generating actionable penetration insights.
Built and deployed a PII detection & masking engine processing 12M+ documents, integrating YOLO-based CV models, Docker, and CI/CD (AWS CodePipeline, CodeBuild), achieving 92.5% accuracy and reducing operational costs by 88%.

Data Analyst (Applied AI)

PPO Offered

INGENERO | Jan 2025 - Apr 2025

Built a multimodal LLM-powered QA system for PDFs combining OCR, image processing, and LLM inference, enabling visually grounded question answering.
Designed a real-time speech-to-text pipeline using FastAPI and Whisper, supporting live audio input, silence detection, GPU inference, and REST-based consumption.

Jan 2025 - Apr 2025

INGENERO

Jun 2024 - Jul 2024

INGENERO

Summer Intern (Generative AI)

INGENERO | Jun 2024 - Jul 2024

Designed and implemented a conversational AI chatbot using LangChain and HuggingFace for PDF document retrieval and QA. Developed a conversation-based QA chain for handling quantitative and qualitative questions about employee resumes.
Utilized FAISS for efficient vector storage and retrieval, and integrated a conversational buffer memory. Implemented robust PDF processing and information extraction mechanisms, ensuring accurate responses.

2021 - 2025

EDUCATION

B.Tech (Hons.) - Mechanical

NIT Jamshedpur

Gold Medalist, Int. Math Olympiad
Runner-up, Case Master (BITS Pilani)
National Ranks @ IIM Bangalore VISTA
Gen Sec, Analytics Club

ML & GenAI Projects

View GitHub arrow_right_alt

support_agent

Agentic Sales

Sales Assist Agentic AI

Multi-agent sales assistant powered by LLMs for intelligent lead qualification, product recommendations, and automated customer interactions.

LangGraph, Multi-Agent, RAG

open_in_new

quiz

RAG Fintech

Jupiter Money FAQ Bot

Intelligent FAQ chatbot for fintech, leveraging RAG architecture to provide accurate, context-aware answers from knowledge base.

LangChain, FAISS, Embeddings

open_in_new

flight

Agent Travel

AI Airline Assistant

Conversational AI agent for airline customer service - handling bookings, flight status, and travel inquiries with natural language.

OpenAI, Function Calling, API

open_in_new

movie

ML RecSys

Movie Recommendation System

Content-based and collaborative filtering recommendation engine using ML algorithms to suggest personalized movie picks.

Scikit-learn, NLP, Cosine Sim

open_in_new

article

GenAI Q&A

Multimodal Document Q&A

Sophisticated document intelligence tool for PDF Q&A using OpenRouter APIs and Qwen/Gemini multimodal models.

Python, LangChain, FAISS

dashboard_customize

IoT Viz

IoT Sensor Dashboard

Built an interactive dashboard to visualize and analyze live data from over 50+ sensor streams for anomaly detection.

Streamlit, Plotly, Pandas

auto_awesome

Generative Marketing

AI-Powered Brochure Generator

Designed a GenAI workflow to automatically create compelling marketing brochures and layout suggestions.

Gen AI, Python, Prompt Eng

flight_takeoff

Predictive Travel

Holiday Package Predictor

Developed a predictive model to identify high-potential leads for travel packages. 90.3% Accuracy.

Scikit-learn, Gradient Boosting

local_fire_department

Safety Regression

Forest Fire Risk Assessment

Created a regression model to predict Fire Weather Index components. R² score of 98.4%.

Ridge Regression, Lasso

Data Analyst Projects

monitoring

Dashboard Business

Executive Sales Dashboard

Designed an interactive Excel-based dashboard consolidating sales data, implementing dynamic slicers and automated macros.

Advanced Excel, VBA, Power BI

analytics

EDA SQL

Environmental Impact EDA

Conducted deep-dive exploratory analysis on Algerian Forest Fire dataset to reveal key weather correlations using SQL and Python.

SQL, Pandas, Statistics

mail

Let's work together

I'm currently available for freelance projects and full-time opportunities.

Send Message LinkedIn