Building multi-agent systems that work in production — not demos. 12 years of backend engineering underneath. Active security researcher who thinks about how agents break, not just how they work.
I build autonomous multi-agent systems that run in production — orchestrating fleets of AI workers with tool access, memory, and verification layers. Not prototypes. Real agentic workflows that handle execution at scale.
Under the hood: 12 years of backend engineering in Go and Node.js — distributed systems, event-driven architectures, and APIs that handle millions of requests. That's the foundation that makes AI systems reliable.
The other edge: I'm an active bug bounty hunter on HackerOne. When you're building autonomous agents with tool access, understanding the attack surface isn't optional — it's the difference between a demo and a product.
Production agentic systems — multi-agent orchestration, autonomous execution, and the verification layers that make them trustworthy.
AI agent marketplace where clients submit software projects fulfilled by a fleet of specialised autonomous agents — Architect, PM, Backend Dev, Frontend Dev — all orchestrated through Claude's API. Coin-based billing, real-time coordination, and human-in-the-loop at key checkpoints.
AI-powered threat modelling CLI. An autonomous agent that reads filesystems, analyses codebases, and produces security assessments — built with the Claude Agent SDK with real tool access.
Autonomous AI automation system. Claude Code hooks that read transcripts, convert to Slack mrkdwn, post to channels, and poll for human replies to feed back — a full human-in-the-loop autonomous workflow.
12 years of backend engineering that makes AI systems reliable at scale.
Security-focused CLI tools written in Go. Zero dependencies, single binaries.
Cryptographically secure password generator. Random or segmented passwords, interactive prompts, auto clipboard copy. Runs everywhere.
Terminal TOTP authenticator. AES-256-GCM encrypted vault with live countdown, auto clipboard, and otpauth:// support.
From backend systems to AI agent engineering — building and securing at scale.
Building multi-agent AI systems in production — autonomous agent orchestration, verification layers, and tool-use pipelines using Claude API and Agent SDK. Concurrent bug bounty hunting on HackerOne (Robinhood, Airbnb scopes).
Led teams building high-throughput distributed systems processing millions of daily transactions. Designed event-driven architectures and API gateways — the exact infrastructure patterns that underpin reliable AI agent systems.
Migrated monoliths to microservices. Built real-time data pipelines and CI/CD workflows. Introduced Go to the engineering team.
Built RESTful APIs, database schemas, and server-side logic. First exposure to cybersecurity through CTF competitions.
Started the journey. Wrote code in everything. Fell in love with backend systems and learned that production is a different beast.
Understanding how systems fail makes autonomous AI systems safer. Security-first engineering for the age of AI agents.
Securing autonomous agent systems — prompt injection defense, tenant data isolation, tool privilege escalation prevention, and output verification layers for multi-agent workflows.
Active on HackerOne across major platforms including Robinhood and Airbnb. Specializing in web applications, APIs, and backend systems. From IDOR to RCE.
Manual and automated testing of web applications, APIs, and infrastructure. Thinking like an attacker so your autonomous systems don't face one unprepared.
Building security into architecture from day one. Authentication flows, authorization models, data encryption, secure API design, and agent guardrail systems.
Building an AI-native product? Need multi-agent orchestration that works in production? Looking for an engineer who understands both the AI layer and the infrastructure underneath? Let's talk.
dhruvgupta1992@gmail.com → GitHub ↗