Chandan Vaishnav

Portfolio

From raw data to decisions people trust.

I build the layer between messy data and action: RAG pipelines wired to the Claude API, async geospatial systems processing satellite imagery through Google Earth Engine, and MLOps workflows that make models reproducible instead of one-off notebooks.

View work Email Chandan

Global sources unified

Satellite monitoring themes

Analyses supported per day

Production ML pipelines

Experience

Data systems with real operating constraints.

Current internship work at SKYQuest Technology Consulting, framed around decision intelligence, geospatial risk, async pipelines, and evidence-backed AI output.

Meridian

AI-powered decision intelligence platform

Built as a 5-layer system connecting data ingestion, a RAG context engine, Claude API sector reasoning, and automated output generation for PPT, Excel, and PDF deliverables.

RAG Claude API 47+ sources Output automation

GCAIP

Geospatial climate-risk decision support

Architected an async FastAPI, Celery, and Redis pipeline for Google Earth Engine analyses with SSE progress streaming, quota-aware caching, and mandatory confidence scoring on every output.

FastAPI Celery Redis Earth Engine SSE

Selected Work

Projects that prove the stack.

A recruiter-friendly view of the systems most aligned with AI engineering, machine learning, geospatial data products, and production deployment.

Flagship SKYQuest / Confirm before publishing

Meridian Decision Intelligence Platform

RAG + Claude API system that turns 47+ global sources into cited, board-ready strategy outputs. The emphasis is grounded context, evidence-backed reasoning, and automated output generation rather than a loose chatbot layer.

02 Geospatial Systems

GCAIP Climate-Risk Pipeline

Satellite-derived monitoring system for flood, erosion, reservoir, and mangrove signals, supported by async job orchestration, confidence scoring, caching, and live progress streaming.

03 MLOps / NLP

YouTube Comments Analysis Chrome Plugin

Real-time comment sentiment classification plus LLM theme summarization, tracked through MLflow, DVC, Optuna, Docker, GitHub Actions, and AWS deployment workflows.

04 Regression / API

SwiggySense Delivery Time Prediction

Delivery latency prediction pipeline with DVC-tracked experiments, a Dockerized REST API, and CI/CD deployment flow for reproducible real-time inference.

05 ML Fundamentals

Student Performance Indicator

Modular ML pipeline for ingestion, transformation, model training, logging, and Flask-based inference, comparing Linear Regression, Lasso, Ridge, and CatBoost.

Skills

A practical stack for intelligent products.

Explicit tooling for recruiters and technical reviewers: the stack is built around production data systems, not portfolio decoration.

Languages & Data

Python, SQL, C/C++, NumPy, Pandas, Statistics, MySQL

ML & AI

Scikit-learn, XGBoost, CatBoost, OpenCV, YOLOv7, LLM integration, RAG, prompt engineering

MLOps & Cloud

MLflow, DVC, Optuna, Docker, GitHub Actions, AWS, Oracle Cloud Infrastructure

Systems & Delivery

FastAPI, Flask, Django REST Framework, Celery, Redis, SSE, Google Earth Engine, Streamlit

About

Calm execution for messy, high-context data problems.

I like problems where the data is messy, the constraints are real, and the output has to be something a non-technical stakeholder can trust in minutes, not weeks. That is the thread across a RAG-based decision platform, a climate-risk geospatial system, and shipped ML products with reproducible pipelines behind them.

Currently: Data Science Intern @ SKYQuest
Education: B.Tech, AI & ML - ITS Engineering College
Credentials: OCI Data Science Professional, Anaconda, Django REST APIs

Contact

Have a sharp data problem worth building?

Send the dataset, the decision bottleneck, or the "we have 47 sources and no time" problem. I can help turn it into a system that is structured, evidence-backed, and fast enough that people actually use it.

Email Chandan GitHub LinkedIn Resume

Please confirm the preferred email, public project links, and SKYQuest disclosure boundaries before publishing.