Skip to main content
Director of AI/ML at FlexTecs

Jayavelu
Balaji

Director of AI/ML at FlexTecs, leading AI-driven innovation in recovery audit and payment accuracy. Building secure, production-grade AI agent systems with expertise in LLM security, agentic AI frameworks, and open-source contributions to LangChain and LlamaIndex.

AI Leadership at FlexTecs

As Director of AI/ML at FlexTecs, I lead the AI and machine learning strategy for a company that reviews over $1 trillion in client transactions and recovers or prevents over $1 billion in incorrect payments annually. I'm building intelligent systems that accelerate recovery audits, strengthen payment accuracy, and power FlexTecs' proprietary SaaS platform, FlexTrap.

Beyond my work at FlexTecs, I contribute to open-source AI frameworks including LangChain, LlamaIndex, and Hugging Face, focusing on security hardening and architectural improvements for agentic AI workflows.

My research on Model Context Protocol (MCP) security has identified 11 critical vulnerability classes affecting enterprise AI deployments, including tool poisoning attacks and cross-server context abuse in agentic systems.

Core Expertise

AI Strategy
LLM Security
Agentic AI
Recovery Audit AI
LangChain
LlamaIndex
Python / ML Ops
Open Source

Projects & Contributions

Enterprise AI at FlexTecs and open-source contributions in AI security.

AI/ML at FlexTecs

Leading AI strategy and ML engineering at FlexTecs, building intelligent systems that power recovery audits across $1T+ in annual client transactions. Driving AI innovation for FlexTrap, the company's proprietary SaaS platform for payment accuracy and error prevention.

AI Strategy FinTech Recovery Audit
Visit FlexTecs

MCP Security Research

Deep technical analysis of 11 critical vulnerability classes in Model Context Protocol implementations, including CVE-2025-6514 (CVSS 10.0), tool poisoning, and cross-server exploitation vectors.

Security MCP Agentic AI
Read Research

LangChain & LlamaIndex

Open-source contributions focusing on security hardening for agentic workflows, tool validation, chain-of-trust mechanisms, and prompt injection mitigation in RAG systems.

LangChain LlamaIndex Open Source
View on GitHub

Hugging Face Integrations

Contributing to model safety tooling and integration patterns for deploying open-source LLMs in enterprise environments with proper security guardrails.

Hugging Face LLM ML Ops
View on GitHub

Articles & Research

Technical deep-dives on AI security, agentic systems, and LLM infrastructure.

Loading articles…

Showing latest 5 articles from Dev.to & LinkedIn.

View All Articles

Let's Work Together

I'm always interested in discussing AI strategy, enterprise ML challenges, collaboration on open-source projects, or the future of AI in financial technology and recovery audit.

Fill in the form and I'll get back to you directly.