IEEE YESIST12 IEngage Track Problem Statements:
1. AI-Driven Cloud Ops (Proactive Anomaly Detection & Self-Healing Cloud Systems)
This problem statement addresses the limitations of current cloud operations models that rely heavily on manual log analysis, rule-based alerting, and human-driven incident response. As cloud environments scale across multi-cloud, microservices, and containerized architectures, fragmented telemetry and lack of automated remediation increase Mean Time to Detect (MTTD) and Mean Time to Resolve (MTTR).
Key Focus Areas:
-
AI-driven anomaly detection across logs, metrics, traces, and security telemetry.
-
Cross-domain event correlation and intelligent incident classification.
-
Automated remediation workflows and self-healing mechanisms.
-
Reduction of operational, compliance, and security risks.
For complete technical scope, objectives, constraints, KPIs, and expected outcomes, refer to the detailed document below:
2. Quantum-Safe Finance (Hybrid Classical–Quantum Security Architecture)
The rapid advancement of quantum computing poses a significant threat to traditional cryptographic systems such as RSA and ECC, which secure financial transactions and sensitive data. Algorithms like Shor’s and Grover’s may enable future decryption of currently encrypted data (‘harvest-now, decrypt-later’ attacks).
Key Focus Areas:
-
Assessment of quantum threats to financial systems and encryption standards.
-
Evaluation of Post-Quantum Cryptography (PQC) algorithms.
-
Design of hybrid classical + quantum-resilient security architecture.
-
Development of a phased migration roadmap for financial institutions.
For complete methodology, architecture design, strategic outputs, and implementation roadmap, refer to the detailed document:
3. Smart E-Waste Platform (Sustainable E-Waste Tracking & Recovery Platform for India)
India is the third-largest producer of e-waste globally, yet lacks a transparent, technology-driven, end-to-end digital ecosystem that ensures traceability, accountability, and incentivized participation across the e-waste value chain.
Key Focus Areas:
-
Digital tracking of electronic products from manufacturing/import to disposal.
-
Integration of formal and informal recycling sectors.
-
Compliance with Extended Producer Responsibility (EPR) regulations.
-
Real-time analytics for regulators and consumer incentive mechanisms.
For detailed regulatory context, constraints, KPIs, and expected outcomes, refer to the detailed document:
1. AI-Assisted Ticket Classification, Urgency Estimation, KB Retrieval, and Grounded Response Drafting
This problem focuses on building a system that classifies ticket category (Refund, Login, Delivery, Billing, Account, or Other), estimates urgency (High, Medium, or Low), retrieves relevant Knowledge Base snippets, and drafts a grounded reply. The system ensures that draft replies are based on approved content and asks for clarification instead of guessing when information is missing.
Key Focus Areas:
-
Automated ticket categorization and urgency estimation.
-
KB ingestion and semantic search for relevant snippet retrieval.
-
Generation of grounded draft replies with source citations.
-
Reduction of agent triage time while ensuring professional, policy-aligned responses.
For complete technical scope, objectives, constraints, KPIs, and expected outcomes, refer to the detailed document below:
2. AI-Based Duplicate Bug Detection, Defect Clustering, Missing Field Analysis, and Report Improvement
Software teams often receive multiple defect reports describing the same underlying issue. Duplicate defects increase triage effort, fragment engineering discussions, and delay resolution. This problem focuses on building a system that detects duplicate or possible duplicate defect reports using similarity search and clustering, assigns cluster IDs, suggests missing fields, and generates an improved defect summary while marking duplicates only when similarity exceeds a defined threshold.
Key Focus Areas:
-
Duplicate defect detection and similarity-based clustering (DBSCAN/Nearest-cluster).
-
Missing field analysis (reproduction steps, logs, environment, etc.) for report enhancement.
-
Embedding generation and vector database based similarity search.
-
Decisioning into duplicate, possible_duplicate, or new_defect based on thresholds.
For complete technical scope, objectives, constraints, KPIs, and expected outcomes, refer to the detailed document below:
3. Grounded Question Answering over Enterprise Documents using Retrieval-Augmented Generation
Organizations maintain large collections of policies, FAQs, manuals, and operational documents. This problem focuses on building a Retrieval-Augmented Generation (RAG) based Q&A bot that ingests documents, retrieves relevant sections, and generates concise answers strictly from retrieved content with citations. If the answer is not present, the system must avoid guessing and ask for the relevant document.
Key Focus Areas:
-
Document ingestion, chunking, and preprocessing for effective retrieval.
-
Embedding generation and semantic search using vector databases.
-
Answer generation strictly grounded in retrieved content with citations (source, snippet, score).
-
Fallback handling for unsupported questions to prevent hallucinations.
For complete technical scope, objectives, constraints, KPIs, and expected outcomes, refer to the detailed document:
