AYUSH MISHRA

I build AI systems that connect natural-language requirements to engineering designs, configurable models, and executable workflows. As Head of Artificial Intelligence at Engineering Intent Corporation, I own AI strategy and hands-on delivery for kBridge assistants, domain-specific model training, and engineering automation.

My work spans the full delivery cycle: customer discovery, prototypes and pilots, solution architecture, implementation, validation, and rollout. I bring 10+ years across enterprise solutions, engineering automation, and technical delivery.

Recent work includes:

• Domain-specific LLM fine-tuning across DeepSeek and Qwen using PyTorch, Hugging Face Transformers, PEFT, TRL, and LoRA/QLoRA.

• Instruction dataset curation and programmatic validation of generated design/project JSON, with bounded generation–validation–retry workflows.

• AI architect workflows that generate kBridge design definitions, JavaScript rules, configurable geometry, and UI components.

• n8n orchestration and MCP tooling for model calls, API operations, session handling, and structured platform interaction.

• Production Salesforce–Oracle integrations, including CPQ, to support quoting, pricing, approvals, and data exchange.

I enjoy working close to customers and engineering teams: turning ambiguous requirements into practical solutions, debugging failures, and making complex workflows easier to use and maintain.

Core tools: Python, JavaScript, C#, REST APIs, SQL Server, n8n, MCP, RAG, LLM function calling, PyTorch, and Hugging Face.

I welcome conversations about applied AI, solutions architecture, and engineering automation.