I'm Vic, an engineer and builder creating the programmable intelligence layer across decentralized protocols, data networks, and autonomous AI systems.

My north star: ship 1B+ USD worth of open systems that accelerate global autonomy, intelligence, and transparency — at scale.

I am heavily involved in optimizing network performance, building scalable data pipelines, and solving complex infrastructure problems across decentralized protocols, agentic systems, and distributed networks.

Presently: Substrate, Solana, EVM, Rollups, Layer2, Agentic AI, LLM Infrastructure, Data Infrastructure, P2P Networking, Node Architecture, Local-First Systems, and Compute Over Data

• 🧱 15 years in software engineering
  • 🌱 15 years of Nodejs
  • 🐍 12 years of Python
  • 🦀 8 years of Rust
  • 🐹 6 years of Go
  • 🏎️ 4 years of C
  • 🦎 Can't stop Zigging
• 🧪 12 years in data, machine learning, and AI
  • 🗄️ 12 years of SQL/NoSQL
  • 🧱 8 years of Big Data
  • 🤖 6 years in AI
• 🕸️ 5 years building Web3 protocols
  • 💠 4 years Ethereum (EVM)
  • 🔗 3 years Substrate
  • ⚡️ 2 years Solana
  • ⛏️ 1 year Bitcoin

{
  "blockchain": {
    "Protocols": ["Ethereum", "Solana", "Polkadot", "Optimism", "Optimism", "Bitcoin"],
    "Clients": ["Substrate", "op-geth", "Firedancer"],
  },
  "agentic_ai": {
    "Frameworks": ["OpenClaw", "Hermes"],
    "Coding Agents": ["Claude Code", "OpenAI Codex"],
    "Models": ["Gemini", "Qwen", "Llama"],
    "Inference": ["vLLM", "Ollama", "LM Studio"],
    "Tools": ["MCP", "RAG"]
  },
  "llm_training": {
    "Model Adaptation": ["LoRA/QLoRA", "PEFT", "Hugging Face Transformers"],
    "Testing & Evaluation": ["Evaluation Datasets", "Benchmarking", "Experiment Tracking"]
  },
  "machine_learning_and_data_science": {
    "deep_learning": ["PyTorch", "TensorFlow"],
    "computer_vision": ["Edge Detection", "Segmentation", "Object Tracking"],
  },
  "data": {
    "Databases": ["PostgreSQL", "MongoDB", "Cassandra", "HBase", "Redis"],
    "Search & Vectors": ["Elasticsearch", "Qdrant", "FAISS"],
    "Graph": ["JanusGraph", "Neo4j"],
    "Streaming": ["Kafka", "NATS"],
    "Big Data": ["Spark", "Flink", "Hadoop"],
    "Warehouse & OLAP": ["ClickHouse", "BigQuery", "Snowflake"],
    "Pipelines": ["Airflow", "Dagster", "dbt"]
  },
  "backend": {
    "Python": ["FastAPI", "Quart"],
    "Node.js": ["Fastify", "NestJS", "tRPC"],
    "Rust": ["Tokio"],
    "Java": ["Spring Boot"]
  },
  "devops_and_monitoring": {
    "Infra": ["Docker", "Kubernetes", "Terraform", "ChaosMonkey", "Jepsen"],
    "Metrics & Monitoring": ["Prometheus", "Grafana", "InfluxDB"],
    "Logs & Search": ["Loki", "Elasticsearch", "Kibana"],
    "Tracing": ["Tempo", "OpenTelemetry"],
  },
}

• tickoni — High-performance AI harness for agentic finance
• cere-network — Decentralized AI Agent Platform
  • payout — Payout, fees, and debt management
  • offchain-worker-etl — Consensus-based aggregation of distributed activity
  • merkle-proofs - Proof-of-activity batching via MMR trees
• op-bnb — OpStack-based optimistic rollup for BNB Chain
  • op-batcher — Foundational contributions to Optimism client to solve race condition
  • op-perf — Node sync improvements and state verification
  • op-node — Client patch for sequencer backpressure handling
• kermes — Solana Restaking Platform
  • Multi-Asset Staking: SPL tokens, SPL-2022 tokens, NFTs, and LSTs
  • Vault Share Tokens: LP-like representation of user shares
  • Reward System: Rewards in tokens, points, or NFTs
  • Secure CPI: Safe cross-program interactions
• random-pedersen — Pedersen commitment and randomness tooling
  • Secure Decentralized Randomness: Pedersen + MPC for tamper-proof RNG.
  • Homomorphic Aggregation: Combine values without revealing inputs.
  • Collaborative Commit-Reveal Protocol: Multi-node process with integrity checks.
  • Axum-based JSON AP*: Minimal HTTP API with ephemeral cache.
• llama-eats — AI-powered food ordering with contextual LLM recommendations
  • LLM-Powered Intent Parsing: Uses quantized LLaMA-2 for dynamic dialogue state tracking.
  • Vector-Based Preference Matching: Embedding models + cosine similarity for dish retrieval.
  • Numerical Inference via Fine-Tuned GPT-2: Parses budget constraints from natural expressions.
  • Context-Preserving Session Embeddings: UUID-based context store for coherent multi-turn ordering.
• raceme-js — Modular clustering framework with genetic algorithm extensions
  • Genetic Clustering Extensions: Implements genetic-based algorithms tailored for social network community detection.
  • Multi-Store Adaptability: Supports planned adapters for Cassandra, MongoDB, and Neo4j.
  • Algorithm Diversity: Supports K-Medoids, Fuzzy C-Means, and Betweenness clustering.
  • Performance & Comparisons: Framework benchmarking and comparing clustering models.