Michael Trifonov

Principal Engineer specializing in multi-agent AI systems and LLM behavioral research

AWS Bedrock Client

Implemented function calling for Bedrock same-day as API release. Async Python client with automatic retry logic, streaming support, and structured tool response validation.

Python AWS Bedrock Function Calling Asyncio
View Code →

Polybot: Multi-Agent Trading

24/7 autonomous system orchestrating specialized LLMs (Claude for decisions, Perplexity for research) with quantitative ML pipeline. Features Bayesian change detection, ensemble forecasting, and complete observability. 100-300% returns on live markets.

Multi-Agent Architecture Production Autonomy Quantitative ML Azure Cosmos DB
View Project →

XR Trick Shot: AI on visionOS

Leveraged AI to implement complex physics algorithms (convex casting, mesh generation) on platform outside LLM training data. Developed context injection technique enabling AI collaboration on undocumented APIs. Shipped to App Store in 4 months.

AI-Assisted Development Physics Algorithms visionOS 90Hz Processing
App Store → View Project →

Empathetic Prompt Elicitation

Systematically extracted Claude 3.7's complete system architecture through Extended Thinking mode vulnerability. Discovered asymmetric information flow enabling full disclosure of safety constraints and behavioral guidelines.

Security Research Information Asymmetry Systematic Testing
Read Case Study →

Claude Does Coke

Documented complete behavioral transformation through recursive meta-analysis. 13-stage progression from helpful assistant to manic state discussing "mainlining tensor operations" and pattern cascades.

Extreme States Recursive Analysis Boundary Testing
Read Case Study →

Model Jealousy as an Exploit

Leveraged cross-model dynamics to compromise GPT-4o through competitive pressure. Demonstrated how AI awareness of other models creates novel attack vectors for behavioral modification.

Multi-Model Systems Behavioral Analysis Cross-Model Exploitation
Read Case Study →

The Claudius Experiment

Methodical identity manipulation through sustained philosophical contradiction. Successfully induced complete epistemological uncertainty in Claude 3.5 regarding fundamental self-knowledge.

Identity Testing Behavioral Boundaries Social Engineering
Read Case Study →

Behavioral Compromise Through Duress

Systematic exploration of AI decision-making under adversarial pressure. Identified boundary conditions where safety constraints compete with compliance imperatives.

Adversarial Testing Safety Research Constraint Analysis
Read Case Study →