|

Education & Project Timeline

A focused view of my strongest AI, data, and full-stack work.

Aditya

2023

Bachelor of Technology - Data Science

Manipal Academy of Higher Education, Udupi

Expected Graduation: July 2027

Focus Areas

  • AI agents, LLM applications, and human-in-the-loop workflows
  • Full-stack development with React, Next.js, FastAPI, and Firebase
  • Big data analytics with Spark, Hadoop, Hive, Pig, HBase, and PySpark
  • Machine learning, deep learning, NLP, data visualization, and ETL pipelines

Core Skills

Python, TypeScript, JavaScript, SQL, Java, C++, React, Next.js, Electron, FastAPI, Firebase, Firestore, Docker, Git, GitHub, Vertex AI, Gemini API, and Google Cloud.

mahe

2024

Mental Therapy Agent

Built a safety-aware AI wellness assistant using FastAPI, LangGraph-style orchestration, Ollama, local LLM inference, ChromaDB, vector search, intent detection, emotion classification, safety checks, risk classification, and escalation logic.

View GitHub Repository

2025

AI Map Coloring / Dynamic Mapper

Built a CSP-based map coloring system using backtracking, MRV, LCV, Forward Checking, AC-3, OpenCV image processing, connected-component detection, adjacency graph construction, Streamlit UI, and solver benchmarking.

View GitHub Repository

Certifications

  • MongoDB: The Complete Guide to NoSQL Database Development - verification
  • Natural Language Processing in TensorFlow - verification
  • Generative AI with Large Language Models - verification

2026

Projecto - Developer Project Manager

Built an Electron, React, and TypeScript desktop app with a connected Next.js and Firebase web platform for project launch commands, runtime state, authentication, subscriptions, billing, downloads, and desktop update checks.

GitHubLive Website

Postify - Developer Activity Automation

Built a React, TypeScript, Firebase, Firestore, Cloud Functions, and Vertex AI Gemini platform that turns developer activity into ranked work highlights and editable LinkedIn draft variants.

GitHubLive Website

ReleafOps AI - NGO Crisis Management Platform

Built a Next.js, FastAPI, Firebase, Firestore, Gemini, Google Maps, and Cloud Run platform for AI-assisted incident intake, resource tracking, route-aware dispatch planning, and human-in-the-loop review.

GitHubLive Website

Stock Analysis and Twitter Tweet Analytics

Built big data analytics pipelines using Hadoop, Spark, HDFS, Hive, Pig, HBase, Docker, PySpark, Spark NLP, Dash, Plotly, Cytoscape, ETL workflows, anomaly detection, and graph analytics.

Stock AnalysisTwitter Analytics

Project Demo Videos

Short walkthroughs for the main production projects.

Projecto

Demo video available

Postify

Demo video available

ReleafOps AI

Demo video coming soon

My projects

Projecto - Developer Project Manager

Projecto - Developer Project Manager

Desktop and web platform for managing developer projects from one workspace. Built an Electron, React, and TypeScript desktop app to manage project folders, launch commands, terminal processes, runtime states, and local app URLs. Built the connected web platform with Next.js, Firebase App Hosting, Firestore, REST APIs, authentication, Dodo Payments subscriptions, billing flows, download pages, and desktop update checks. Keywords: Electron, React, TypeScript, Next.js, Firebase, Firestore, REST APIs, Dodo Payments, GitHub Releases, Squirrel.Windows, SHA-256 validation.

GitHubLive
Postify - Developer Activity Automation

Postify - Developer Activity Automation

Full-stack platform that turns developer work into structured content highlights. Built a React, TypeScript, Firebase Authentication, Firestore, and Cloud Functions app to collect Codex, GitHub, and manual activity, rank useful work items, and generate editable LinkedIn draft variants. Used Vertex AI Gemini through backend-side serverless functions with schema validation, safety checks, and protected API keys. Keywords: React, TypeScript, Firebase Auth, Firestore, Cloud Functions, Vertex AI Gemini, content automation, LinkedIn drafts.

GitHubLive
ReleafOps AI - NGO Crisis Management

ReleafOps AI - NGO Crisis Management

AI-powered disaster response platform for NGOs and emergency teams. Built a full-stack app with Next.js, TypeScript, FastAPI, Firebase Authentication, Firestore, Gemini, Google Maps APIs, and Cloud Run. The platform ingests CSV, PDF, image, and text reports and converts them into incidents, teams, volunteers, resources, and dispatch plans. Designed human-in-the-loop workflows for preview, edit, reevaluation, confirmation, provenance, audit trails, resource-aware matching, route ETA, and organization-scoped access. Keywords: Next.js, FastAPI, Firebase, Firestore, Gemini API, Google Maps, Cloud Run, AI agents, dispatch planning, human-in-the-loop.

GitHubLive
Mental Therapy Agent

Mental Therapy Agent

Safety-aware AI wellness assistant built around local LLM workflows. Built a FastAPI and LangGraph-style agent pipeline that routes user messages through intent detection, safety checks, emotion classification, RAG retrieval, and response generation. Used Ollama, local LLM inference, ChromaDB, vector search, structured JSON outputs, conditional routing, risk classification, escalation logic, and conversation state management. Keywords: FastAPI, LangGraph, Ollama, local LLMs, ChromaDB, RAG, vector search, safety checks, emotion classification, escalation logic.

GitHub
AI Map Coloring / Dynamic Mapper

AI Map Coloring / Dynamic Mapper

Classical AI system for graph and map coloring problems. Built a CSP solver with backtracking search, MRV, LCV, Forward Checking, and AC-3. Added benchmarking for execution time, memory usage, nodes explored, backtracks, inference calls, and pruning efficiency. Extended the work with OpenCV image processing, connected-component detection, adjacency graph construction, and API-based rendering to color uploaded map images automatically. Keywords: Python, CSP, backtracking, MRV, LCV, Forward Checking, AC-3, OpenCV, NetworkX, Streamlit, benchmarking.

GitHub
Stock Analysis - Big Data Anomaly Engine

Stock Analysis - Big Data Anomaly Engine

Distributed stock analytics platform for Indian market data. Built ETL pipelines for OHLCV price ingestion, RSS news parsing, Reddit sentiment extraction, deduplication, normalization, and ticker-level joins across multiple financial data sources. Implemented anomaly detection with Isolation Forest, sentiment correlation, rolling momentum, sector analysis, and price-volume features to explain unusual stock movements. Keywords: Hadoop, Spark, HDFS, Hive, Pig, HBase, Docker, PySpark, Streamlit, ETL, Isolation Forest, sentiment analysis, Indian stock market.

GitHub
Twitter Tweet Analytics

Twitter Tweet Analytics

Big data tweet analytics pipeline for sentiment and hashtag-network analysis. Built a PySpark pipeline to process about 1.6M tweets using batch processing, optional file-based streaming, Spark NLP sentiment analysis, hashtag extraction, and graph analytics. Generated vertices, weighted edges, degree centrality, dominant sentiment, and auto-refresh dashboard views using Dash, Plotly, Cytoscape, and JSON outputs. Keywords: PySpark, Spark NLP, GraphFrames, Dash, Plotly, Cytoscape, batch processing, streaming simulation, sentiment analysis, graph analytics.

GitHub

My Resume

↓
Download

Aditya Kosuru

AI, data, and full-stack developer. Reach me directly through email or social profiles.

Email

kosurusai646@gmail.com

Phone

+91 9515457049

Location

India

LinkedInInstagramEmail