Distributed Systems · Cloudflare Edge · AI Agents
Building high-throughput distributed backends & agent runtimes.
Deploying distributed systems across Cloudflare edge infrastructure (Workers, Queues, Hyperdrive), high-frequency C++17 trading engines, and autonomous agent compiler toolchains. Creator of AgentIR and CodePlanner CLI.
Live Systems & 3D Sandboxes
Execute client-side AST compilations, simulated multi-tenant DAG flows, and microsecond order book reconstructions.
AgentIR
Much like LLVM abstracts machine instruction sets from programming languages, AgentIR abstracts agent architectures (prompts, tools, state graphs, handoffs, guardrails) from proprietary framework SDKs. Write once and compile to idiomatic LangGraph, Agno, OpenAI Agents SDK, CrewAI, or MCP with zero runtime latency.
{
"$schema": "https://agentir.dev/v1/spec.json",
"name": "ProductionSecurityCoordinator",
"state": {
"audit_findings": { "type": "list", "reducer": "append" },
"threat_level": { "type": "enum", "values": ["LOW", "HIGH", "CRITICAL"] }
},
"coordinator": {
"model": "gemini-2.5-flash",
"teams": ["StaticAstAnalyzer", "CveVulnerabilityScanner"],
"routing_strategy": "dynamic_confidence"
}
}# Target: LangGraph (StateGraph)
from typing import TypedDict, Annotated, List
from langgraph.graph import StateGraph, END
import operator
class CoordinatorState(TypedDict):
audit_findings: Annotated[List[dict], operator.add]
threat_level: str
def route_next_agent(state: CoordinatorState) -> str:
return "cve_scanner" if state["threat_level"] == "CRITICAL" else "ast_analyzer"
graph = StateGraph(CoordinatorState)
graph.add_node("coordinator", triage_node)
graph.add_node("ast_analyzer", static_ast_node)
graph.add_node("cve_scanner", cve_scanner_node)
graph.add_conditional_edges("coordinator", route_next_agent)
app = graph.compile()Cloudflare Edge & High-Throughput Ingestion
Production distributed topology: V8 isolate edge routing, queue decoupling, global database connection pooling, and Zero Trust cluster ingress.
Cloudflare Workers & Pages
Distributed V8 isolate edge runtime handling incoming traffic, WebSocket duplex connections, and JWT validation across 300+ cities globally with zero cold starts.
// Cloudflare Worker Edge Gateway (TypeScript)
export default {
async fetch(request: Request, env: Env): Promise<Response> {
const url = new URL(request.url);
// 1. Edge Authentication & Fast Path KV Check
const token = request.headers.get("Authorization");
const session = await env.SESSIONS_KV.get(token);
if (!session && !url.pathname.startsWith("/public")) {
return new Response("Unauthorized", { status: 401 });
}
// 2. WebSocket Upgrade for Real-time Streaming
if (request.headers.get("Upgrade") === "websocket") {
const pair = new WebSocketPair();
return handleSessionSocket(pair[1], env);
}
// 3. Dispatch to Cloudflare Queue for async processing
await env.WORKFLOW_QUEUE.send({
event: "task.submitted",
payload: await request.json()
});
return Response.json({ status: "ENQUEUED_AT_EDGE" }, { status: 202 });
}
};Engineering Track Record
Career Experience
Engineering history across distributed pipelines, Cloudflare edge, and agentic platforms.
- Engineered graph-based behavioral intelligence models in Memgraph with causal relationships attached during high-throughput ingestion.
- Implemented belief-state validation and trust dynamics for autonomous agent workflows and decision auditing.
- Designed multi-agent orchestration on Agno with coordinator-to-persona routing and dynamic context retrieval backed by Qdrant.
- Restructured agent memory across PostgreSQL, Redis session storage, and vector retrieval layers.
- Engineered production Bun/Hono → Kafka → FastAPI event streaming pipelines and managed AWS cloud infrastructure across EC2 and RDS.
Systems & Architecture Stack
Production-tested technologies across edge compute, distributed messaging brokers, in-memory graph theory, and compiler pipelines.
Cloudflare & Edge Infrastructure
Deploying and scaling serverless edge systems globally
Backend & Systems Architecture
Building scalable, distributed, and event-driven services
AI Agents & Autonomous Systems
Designing multi-agent orchestration, context, and compiler toolchains
Databases & Graph Intelligence
Relational modeling, graph theory, and time-series data stores