AI Auditor & Content Engineer

Tundra Technologies

Core member of the AI audit team responsible for the systematic evaluation of LLM outputs within high-volume production environments. I collaborate on building robust quality frameworks to detect stylistic drift, ensure factual alignment, and implement hallucination mitigation strategies across the production pipeline.

• Contribute to the identification and correction of failure patterns to ensure consistent model reliability for 10M+ user environments.
• Participate in the development of systematic quality frameworks to standardize AI performance benchmarks.
• Analyze production-level outputs to ensure LLM responses remain grounded in verifiable truth.

Mercor

Contributed to advanced AI model training and evaluation across multiple modalities including text, vision, and audio. Specialized in adversarial testing, reasoning fidelity assessment, and systematic quality evaluation for production-grade AI systems.

Innodata

Implemented a systematic QA process to evaluate LLM responses for hallucinations and stylistic drift, identifying and correcting failure patterns before they reached production. Designed complex, source-grounded prompts to maintain high-fidelity output and brand voice consistency across enterprise-scale datasets, ensuring every response remained grounded in a verifiable source of truth.

• Specialized in reducing hallucination rates through iterative testing and strict "Source Anchor" constraints to ensure factual alignment.
• Optimized the coherence and readability of AI-generated content through systematic prompt engineering and technical auditing.
• Developed internal QA guidelines for evaluating model fidelity and stylistic consistency for Fortune 500 deliverables.

Welocalize

Built systematic evaluation frameworks and scalable annotation guidelines for LLM pipelines. Led cross-functional teams in developing data quality standards and code review processes for generative AI projects.

Drips & OnlyDust

Active open-source contributor with 1,350+ all-time points and 8 high-complexity issues resolved across the Drips, Stellar, and Soroban ecosystems. Specialized in building secure middleware and auditing contract logic for decentralized protocols.

• Developed Stellar wallet authentication systems using SEP-10 protocol and cryptographic nonce management for Traqora.
• Implemented Contract State Verification and specialized Audit Functions to ensure cross-protocol integrity.
• Defined Soroban Resource Cost Models and implemented Global Exception Filters for robust error handling.