AI/ML Engineer | Verifiable AI, Agent Evaluation & Systems Architecture
first conversation is free, sign up to message Alexander
I review AI-assisted or "vibe-coded" apps, identify bugs and broken user journeys, write a structured report, and implement priority fixes to get the product closer to launch. Includes QA notes, repro steps, severity ratings, fix recommendations, and hands-on fixes where needed.
I test core user flows end-to-end and document bugs with expected vs actual behavior, severity, screenshots/notes, and recommendations. Best for founders preparing demos, launches, investor reviews, or accelerator demo days.
I build production-minded AI systems that prioritize reliability, verifiability, and transparent audit trails. My work includes LLM orchestration, multi-agent systems, adversarial evaluation, deterministic quality checks, and AI red teaming. I create systems that validate model outputs against explicit rules, live data sources, test suites, and reproducible evidence rather than relying on unverified AI responses. I have built AI security triage workflows for cross-chain DeFi, on-chain verifiable reasoning systems, multi-engine transcription QA tools, healthcare voice-agent simulators, and compliance-aware grant-drafting agents. My technical stack includes Python, TypeScript, Solidity, Bash, Splunk/SPL, MCP, Flask, Next.js, EVM tooling, REST APIs, Slack integrations, and SDK design. I also bring hands-on infrastructure experience building, repairing, benchmarking, and diagnosing NVIDIA DGX/HGX A100 and H100 systems. I can help with AI-agent evaluation, red-team testing, deterministic validation pipelines, API integrations, smart contracts, security-focused automation, GPU diagnostics, and reliable end-to-end AI product development.