Akash Meesala

At Deccan AI Experts, I evaluated and validated 1,000+ AI model outputs using structured frameworks to support high-quality training data for machine learning and NLP. I also conducted error analysis, data annotation, and quality assurance, identifying recurring model failure patterns.

I designed and deployed V.I.R.U.S., a private multimodal AI desktop assistant using Groq-hosted models for vision, text, and voice. Its features include real-time web search, PDF parsing, voice-based text-to-speech, automated reminders, and daily briefings.

For my AI Smart City project, I built a grievance platform combining YOLOv8 visual defect detection with a LangGraph multi-agent workflow for complaint categorization, SLA scoring, and work order dispatch. The project also includes a municipal copilot with live database grounding, multilingual voice dictation, GIS heatmaps, and ward-level risk forecasts.

I also developed a Telegram chatbot using retrieval-augmented generation and a vector database for context-aware responses. I’m seeking an entry-level role where I can deliver robust, scalable AI solutions.