telecom-ai

PRISM — Photonic Recursive Intelligence with Synaptic Memory · Coherent IQ fiber neural network · Recursive meta-prompt · 4 telecom classifiers at 100% accuracy · IIT Mandi · Himshikhar 2026

Deep learning-based 5G handover prediction using real drive-test measurements, RSRP time-series forecasting, and LSTM/GRU/TCN models for predictive mobility management.

Knowledge Graph-Enhanced LoRA Fine-Tuning for Intelligent 3GPP-Compliant 5G Network Slicing. Maps raw 5G telemetry to 3GPP standards using SVO triples and fine-tuned Llama-3.2-1B models.

Agentic AI-driven framework for 3GPP-compliant (TS 23.501) resource allocation in 5G NR Satellite Networks. Integrates ML models as tools for a LangChain ReAct agent.

Colosseum-based O-RAN slice resource allocation via Federated Learning & Differential Privacy (Opacus). ClusteredFL · FedProx · DQN · PyTorch

Agentic-RAN is a reproducible PyTorch benchmark for comparing neural-network architectures on RAN-oriented regression and sequence-prediction tasks, including MLP, attention, liquid-dynamics, and xLSTM scenarios. It provides Docker-based execution, standardized evaluation metrics, prediction plots, and automated benchmark reports.

Full-stack Generative AI platform that transforms live AI, Telecom and Marketing news into multi-platform social content, branded PDFs and automated email packages using Next.js, FastAPI and Google Gemini.