Building practical AI systems
that ship to production.
I'm Aryan Shukla — an AI/ML engineer building RAG pipelines, anomaly-detection systems, and simulation-driven architectures with Python & FastAPI.
Where AI meets real engineering.
B.Tech Computer Science (AI/ML) student with hands-on experience designing, building, and deploying real ML systems end-to-end. I work where AI meets backend engineering — building FastAPI services, RAG pipelines, anomaly detection models, and simulation-driven architectures. I care about practical impact: shipping projects that run in production, not just notebooks.
Tools I reach for.
A pragmatic stack focused on Python, ML, and shipping APIs that hold up in production.
Projects that actually ship.
Every project below has a clear problem, real implementation, and — where possible — a live deployment.
RAG FastAPI Assistant
Document-aware Q&A API powered by embeddings + FAISS.
- Document ingestion + semantic chunking pipeline
- FAISS vector store with sentence-transformer embeddings
- Modular retrieval and response generation layers
AI Login Anomaly Detection
Behavioral anomaly detection for suspicious login activity.
- Isolation Forest model on engineered behavioral features
- Time, device, and location encoding pipeline
- Threshold-based scoring for real-time flagging
Adaptive Duality Engine
Self-adjusting infrastructure simulator with live metrics.
- Simulated stream of CPU, request rate, active instances
- Adaptive decision engine reacting to live signals
- Modular architecture mimicking production telemetry
RAG FastAPI Assistant
Document-aware Q&A API powered by embeddings + FAISS.
AI Login Anomaly Detection
Behavioral anomaly detection for suspicious login activity.
Adaptive Duality Engine
Self-adjusting infrastructure simulator with live metrics.
V.I.G.I.L.A.N.T
Security-focused monitoring system.
MediAuth (Django)
Authentication-focused medical platform prototype.
Hazelnut ML Classifier
Image / data classification pipeline.
IndiRoadsSim
Simulation of Indian road traffic dynamics.
Traffic Pothole Simulator
MATLAB simulator modeling pothole impact on traffic flow.
ClimateCrowd-TrafficTwin
Digital-twin style model combining climate, crowd, and traffic signals.
EmotionAutoCity
Emotion-aware autonomous city simulation (MATLAB).
Mindcity
MATLAB-based city cognition simulation.
Mood Food
Mood-aware food recommendation prototype.
ZENSORATECH
Task project for ZENSORA.
Built in the open.
Explore my repositories, contributions, and project work on GitHub.
Experience & Education.
Independent AI/ML Developer
- Built FastAPI-based machine learning applications and backend APIs.
- Developed and deployed AI projects involving anomaly detection and document retrieval (RAG).
- Integrated ML models with REST APIs for real-time inference.
- Owned end-to-end delivery: data processing, model integration, deployment, and testing.
B.Tech, Computer Science & Engineering (AI/ML)
United Institute of Technology, Prayagraj
Deepening LLM agent architectures, vector retrieval, and production ML deployment patterns — while shipping projects publicly on GitHub.
Let's build something real.
Open to AI/ML internships, backend roles, and collaboration on practical applied-AI projects.