Design and build agentic AI workflows that generate behavioral and circuit-level simulation models (MATLAB/Simulink, SPICE) for semiconductor devices from technical documentation.

Architect LLM inference across commercial APIs and self-hosted open-weight models served on GPU infrastructure, optimizing for quality, latency, and cost.

Build automated generate, simulate, and refine loops that verify model behavior with deterministic, numerical checks rather than just LLM judgment.

Website: Marylandersonline.umd.edu 

Create Multilingual conversation AI Agent workflow to support the Digital Navigator callcenter, using RAG and MCP Servers, with connections with Twilio and Google Cloud to offer a omnichannel gateway to access it. By leading a small engineering team.

Conducted Research on AI and RAG, LLM models in enhancing digital literacy & equity, identifying multiple high-impact use cases that can benefit the Project, leading to the development of these AI-powered features that improved the learning experience.

Experienced in Incident Response, Threat Hunting, malware, email, and phishing analysis with a track record of timely detection and mitigation of potential threats such as ransomware, APTs, Botnet, DDoS, phishing, and impersonation attacks.  

As a key member of an MSSP, proactively performed 24/7 threat hunting, risk mitigation, and incident response. My contributions helped scale the SOC from 6 clients to successfully overseeing 16 clients within 2 years.

Built an AI-powered B2B FinTech app displaying invoice details & predicts payment dates for businesses.  

Designed and developed an AI-powered B2B FinTech app using Java, Servlet, JS, MySQL, ExtJS, Python, and Scikit-learn that extracts invoice details and predicts payment dates for businesses.

Gained full-stack software engineering skills to complement cybersecurity background.