Principal Full Stack & ML Engineer | AI Research Contractor | 9+ Years of Experience

Mountain View, CA | Email | LinkedIn | GitHub

Elite AI Research Contractor specializing in Large Language Models, Transformer architectures, and production-grade AI systems. Currently working with three of the world's largest AI research laboratories through Mercor, providing CTO-level task orchestration, adversarial testing, and model evaluation for next-generation coding agents.

With 9+ years architecting high-availability platforms and neural interface systems, I bridge abstract AI research with scalable, production-ready execution. Proven track record of 60%+ system latency optimization and 25% user engagement improvements through adaptive AI solutions.

Languages:
Python • C++ • JavaScript/TypeScript • SQL • Ruby • R

AI & Machine Learning:
PyTorch • TensorFlow • Hugging Face Transformers • LLM Fine-tuning • RLHF • Model Evaluation • Adversarial Testing • Vertex AI • OpenCV • EEG Signal Processing

Web & Infrastructure:
React.js • Node.js • Ruby on Rails • GraphQL • Docker • Kubernetes • AWS • GCP

Data & Analytics:
Pandas • NumPy • Apache Spark • Financial Modeling • Algorithmic Trading • Qdrant RAG • Supabase

• Conducting CTO-level evaluation of frontier coding agents and LLMs not yet publicly released
• Performing comprehensive adversarial testing (red-teaming, jailbreak attempts, prompt injection)
• Designing complex coding challenges across environment configuration, database schema, API development
• Generating RLHF/DPO training data for model improvement
• Collaborating with leading AI researchers on evaluation benchmarks

• Architected LLM-based dialogue engine using custom Transformer models for NPC interactions
• Built real-time player behavior analysis pipeline +25% user retention
• Engineered emotionally intelligent game agents with optimized inference latency
• Designed scalable microservices architecture

• Engineering advanced ML models for alpha generation in financial datasets
• 40% improvement in backtesting pipeline execution speed

• Maintained high-traffic platform (10k+ users) achieving 99.9% uptime
• 60% latency reduction through API and microservice optimization
• 40% increase in user engagement via recommendation algorithms

• Engineered firmware and signal processing for EEG wearable devices in C++
• Managed high-throughput biometric data pipeline for 1,000+ participants

Active Research Areas:

• Novel human-AI interaction frameworks
• Transformer architectures for therapeutic dialogue
• Psychological network analysis and personality profiling
• Neural brainwave signal processing

Notable Projects:

• LLM Gameplay Integration - AI-powered gameplay with Claude Vision, Qdrant RAG, Supabase (shadow-of-the-wyrm-AI, crawl_AI forks)
• Neuro-Hiring Platform - Next.js hiring platform with Five-Factor Model personality profiling
• AI-Meta Glasses Integration - Autonomous behavioral event processing for habit tracking
• Open Source Contributions - Open-LLM-VTuber, BasedHardware/omi, roguelike AI projects

• 9+ years architecting enterprise systems
• 60%+ average latency optimization across projects
• 99.9% platform uptime maintained
• 25%+ user engagement improvements
• 1,000+ biometric data participants managed
• 10,000+ concurrent users supported

• 📧 Email: nikita@diannt.net
• 💼 LinkedIn: linkedin.com/in/nikita-annt
• 🐙 GitHub: github.com/diannt

Always exploring the intersection of AI research, human-centered design, and scalable engineering.