Invaya Systems Logo

Invaya Systems

AI Solutions

Advanced artificial intelligence solutions leveraging cutting-edge techniques for knowledge extraction, forecasting, and physics-aware modeling.

Featured Projects

AI-Powered Question Generation with Graph RAG
Created an automated question generation system for US/UK school papers that achieved >90% accuracy while reducing costs from ₹50-₹100 per question to <₹1 per question, leveraging Graph RAG for accurate knowledge retrieval from textbooks.

Challenges

  • ChatGPT reduced cost from ₹50-₹100 to ₹5-₹10 per question but accuracy dropped from 95%+ to ~70%
  • Generating hundreds of questions in a single prompt caused contextual bleed, where earlier outputs influenced later ones
  • Standard vector search retrieved many redundant or overlapping chunks from textbooks
  • This resulted in inconsistent question quality, topic drift, and reduced factual accuracy
  • The drop in correctness made automated generation commercially unviable at scale

Solutions

  • Generated one question at a time to preserve context integrity and improve factual correctness
  • Used three LLMs in sequence: LLM-1 generates question text, LLM-2 generates MCQ options and correct answer, LLM-3 validates for quality
  • Integrated Graph RAG with first and third LLMs for reference-based grounding from specific textbooks, enabling both structured and semantic search
  • Built document-derived knowledge graphs from textbooks where each node represents a concept with vectorized embeddings
  • Applied context engineering, nano models (LLM-1), and prompt caching to minimize token usage
  • Implemented using LangGraph framework with OpenAI GPT-4o as LLM

Outcomes

  • Reduced cost to <₹1 per question, including formatting and direct PDF generation
  • Achieved >90% correctness, validated through human checks for 1 month
  • Improved retrieval accuracy through contextual graph relationships from textbooks
  • Fully automated process eliminated human intervention
  • Saved approximately ₹60,000/month by replacing two manual operators
Satellite Anomaly Detection & Classification
Developed dual-approach system for detecting and classifying anomalies in satellite digital twin telemetry, ensuring estimation accuracy critical for mission safety and cost control.

Challenges

  • Telemetry anomalies (voltage drops, subsystem drifts) caused single-model digital twins to fail
  • Fixed models couldn't capture coupled dynamics (e.g., voltage drop from reaction wheel friction vs battery resistance)
  • Customers relied on digital twin for critical operations like maneuver planning and fuel budgeting
  • Any deviation in digital twin estimations reduced reliability and mission safety

Solutions

  • Applied Kalman Filter-based anomaly detection monitoring deviations between measured and estimated values
  • Introduced latent-space transformation to decorrelate subsystem interactions, making innovation covariance matrix diagonal
  • Implemented LSTM-based Autoencoder for unsupervised anomaly detection, capturing nonlinear temporal dependencies
  • Built 1D CNN classifier trained on Autoencoder residuals to classify fault types using supervised learning

Outcomes

  • Achieved >99% precision and recall on synthetic fault datasets
  • Successfully isolated anomalies for individual subsystems
  • Enabled early fault detection before catastrophic failures
  • System ready for deployment pending real fault telemetry data validation

Ready to Transform Your Business?

Let's discuss how our solutions can solve your specific challenges.