Sudarshan B

About Sudarshan B

I’m an AI Quality Assurance Engineer specializing in LLM evaluation, Conversational AI testing, and AI Safety validation.My work focuses on ensuring AI systems behave safely, reliably, and without harmful or biased outputs in real-world environments. I test AI not just for functionality, but for risk, fairness, and compliance.At Amazon, I contribute to improving Alexa AI systems and smart devices, conducting adversarial testing, response evaluation, and model risk analysis.Key expertise:• LLM response validation (hallucinations, unsafe content, policy violations)• Conversational AI behavior testing• AI safety & fairness evaluation• Red teaming & adversarial testing• Multilingual AI validation• Regression testing on AI model updates• AI-device interaction testing (Amazon Astro)I’m passionate about building trustworthy and human-safe AI systems.

Experience

1. Ai Quality Assurance & Red Teaming Analyst Amazon

   Amazon

   Oct 2024 — Present · Mahadevpura Bengaluru, IN

   Perform adversarial (red team) testing on Alexa conversational AI to detect unsafe and biased outputsEvaluate LLM responses for hallucinations, policy violations, and model failuresConduct AI safety audits aligned with Trust, Privacy & Safety standardsValidate AI behavior across multilingual locales (US, UK, Germany, Spain, Italy)Identify model defects, improving AI reliability and user trustExecute regression testing after model updatesTest Kids Mode AI scenarios for age-appropriate safetyConduct AI fairness and bias evaluationValidate AI behavior across Alexa apps and physical devices

Education

• New Horizon College of Engineering

  Bachelor of Engineering - BE, Information Science/Studies

• Krupanidhi School of Management - Bangalore, India

  PUC, PCMC

• Sophia Primary High School

  10th