● Open to Work

Raseeca Kashelkar

Full Stack Developer | AI/ML
Mumbai

About Me

I'm an Information Technology student at Thadomal Shahani Engineering College with a focus on AI/ML and full-stack development. What I find interesting isn't just building models it's the entire journey from data to deployment and a complete system someone can actually use. My interests span full-stack development, machine learning, agentic AI, and LLM-powered applications. I enjoy taking ideas from concept to production, whether that means training a model, building the backend around it, or creating the experience that ties everything together. Outside of academics, I spend my time building AI projects, participating in hackathons, contributing to open source, and figuring out what's next in agentic AI.

Skills & Technologies

PythonCC++Core JavaJavaScriptTypeScriptSQLReactNext.jsNode.jsExpressFastAPIFlutterREST APIsVercelShadcn/UIScikit-learnPyTorchMachine LearningComputer VisionNLPAgentic AIPostgreSQLMongoDBFirebaseSupabaseNeonGitGitHubDockerPostmanfull stack devai ml

Work Experience

Web Developer Intern

WEBWORK
Dec 2023June 2024

Developed and maintained responsive web interfaces for client projects, contributing to UI implementation, content management, and feature enhancements while ensuring cross-device compatibility and usability. Collaborated with clients and development teams to conduct software testing, gather requirements, demonstrate solutions, and implement content and database updates to support successful project delivery.

Projects

ScamRadar: AI-Powered Fake Job Posting Detector

Built a hybrid fraud detection system combining TF-IDF + Logistic Regression, a rule-based red-flag engine, and Gemini AI contextual reasoning into a weighted ensemble; achieved 86.23% F1-score, 89.44% precision, and 83.24% recall on 3,576 test samples. Integrated Tesseract OCR for screenshot-based analysis and deployed via FastAPI with explainable risk scoring and plain-language fraud assessment.

ResQAI: Multimodal Disaster Detection System

Built a multimodal disaster classifier (Earthquake, Fire, Flood, Normal) using a weighted CNN ensemble with VGG16 (95.53% accuracy, 0.9554 F1-score), ResNet18 (89.30% accuracy), and a Base CNN (88.57% accuracy) for image classification, alongside TF-IDF + Logistic Regression (92.5% accuracy) for text analysis. Deployed the full multimodal pipeline via Flask REST API with confidence-scored predictions, drag-and-drop image upload, and JSON report export.

ScrapHappens: Circular Economy AI Marketplace

Built an AI-powered circular economy marketplace that enables industrial waste exchange through a custom multi-agent architecture with Supervisor Routing, coordinating Buyer, Seller, and DIY agents for intelligent material discovery and workflow automation. Integrated Gemini AI, Socket.io, Twilio IVR, ShipRocket, and Three.js to deliver real-time agent interactions, voice-enabled material discovery, logistics automation, and gamified sustainability impact tracking.

Bookify: AI-Powered Appointment Management System

Built an AI-powered appointment booking platform integrating LangChain, GPT-4o, OpenAI Whisper, and OpenAI TTS to enable conversational and voice-based appointment scheduling. Implemented Twilio IVR, PhonePe payment integration, a slot reservation system to prevent double bookings, and RBAC across multiple user roles.

Education

Bachelor of Engineering (B.E.)

Information Technology

Thadomal Shahani Engineering College (University of Mumbai)2024 - 2027

Diploma

Diploma in Computer Engineering

Government Polytechnic Mumbai (Autonomous)2021 - 2024
Raseeca Kashelkar - Full Stack Developer | AI/ML | HiringAnt