Principal Product Security Engineer at EigenLabs

Anto Joseph

I lead product and infrastructure security at EigenLabs across protocol design, governance, EigenAI, detection tooling, applied cryptography, and adversarial AI systems.

Current focus: EigenAI threat modeling and protocol security · Interests: protocol security, applied cryptography & ZK, adversarial ML, detection tooling

~$22B
Protocol Capital Protected
12+
Years in Security
2018
Helped Launch AI Village
6
OSS Projects
4
Awards & Recognition

Who I Am

I am a developer turned security researcher focused on systems where software, incentives, and adversaries meet. At EigenLabs, I lead product and infrastructure security across protocol, services, governance, and deterministic AI inference, with mitigations protecting ~$22B in TVL on EigenLayer.

Before EigenLabs, I led security work at Coinbase across the MPC wallet, ETH staking, Coinbase Wallet, and Base L2. My earlier work spans consumer abuse defense at Tinder, mobile and embedded security at Intel and Citrix, and hands-on penetration testing and training at EY.

My research roots include adversarial ML since 2017, applied cryptography and ZK, Android instrumentation, distributed fuzzing, and automated red-teaming. I helped start AI Village at DEF CON in 2018, working on the Call for Villages, CFP paper review, and operations in its first year. I have presented research at DEF CON, Black Hat USA, TROOPERS, HITB, BruCon, Shakacon, Hacktivity, and other security conferences.

01 / Protocols
Economic security at protocol scale
Product and infrastructure security for protocol design, services, governance, monitoring, and operator activity.
02 / Frontier AI
Frontier AI security research
Security research across evals, AI safety and alignment, and the infrastructure and tooling that supports frontier model development.
03 / Digital Assets
Applied cryptography for digital assets
Threat modeling and SDLC for MPC wallet systems, ETH staking, bridge security, and L2 scaling infrastructure.
04 / Research
Adversarial tooling and training
Open-source work and talks across distributed fuzzing, Android instrumentation, adversarial ML, and automated red-teaming.
05 / Autoresearch
Autonomous security research products
Agent-driven research platforms for cryptographic circuit optimization, benchmarked against industry leaders and featured in frontier model launches.
06 / Private Inference
Private distributed inference
Privacy-preserving inference that routes encrypted requests to verified hardware, live in production on a major model routing platform.

Security Domains

Protocol Security
Smart Contracts
Applied Cryptography & ZK
MPC Wallets
Mechanism Design
EigenAI / AI Inference
Adversarial ML
Agentic System Security
Detection Pipelines
Distributed Fuzzing
Mobile & Embedded
Product Abuse Defense

Work Experience

Sep 2023 – Present
EigenLabs
Principal Product Security Engineer
Remote
  • Lead product and infrastructure security across protocol, services, and governance; mitigations defend ~$22B in EigenLayer TVL against adversaries with strong economic incentives.
  • Threat-model secure design of EigenAI deterministic AI inference, including adversarial operator behavior, model and output integrity, and incentive-aligned slashing.
  • Build detection and security automation tooling, including continuous monitoring pipelines and signal aggregation across operator activity.
  • Active in cross-organization security research initiatives including SEAL, DSS, and BSidesSF.
Feb 2021 – Aug 2023
Coinbase
Staff Security Engineer
San Francisco Bay Area
  • Drove SDLC and threat modeling for the MPC wallet and enabled ETH staking for Retail and Prime customers. First project at Coinbase was the Ethereum staking launch; those validators went on to run 3.84M ETH (11.42% of all staked ETH) with zero slashing events since genesis, per Coinbase's 2025 validator report.
  • Researched and shipped mitigations against persistent, financially motivated attackers across smart contracts, DeFi protocols, scaling and privacy primitives, and zero-knowledge systems.
  • Designed bridge-security review processes for major scaling solutions; led security for Coinbase Wallet and Base L2.
  • Co-authored Coinbase Engineering's public Euler Finance exploit investigation and presented "Flash Loans Demystified" at DEF CON 29 Blockchain Village.
Mar 2018 – Feb 2021
Tinder
Senior Application Security Engineer
West Hollywood, CA
  • Performed application security assessments, threat modeling, design reviews, and code review for a high-scale consumer platform handling sensitive user data.
  • Designed product-level mitigations against coordinated abuse, impersonation, and harassment vectors.
  • Contributed to mobile security tooling, including open-source contributions to MobSF, and presented FuzzCube at Hacktivity 2020.
  • Helped start AI Village at DEF CON 26 (Call for Villages, CFP paper review, and operations). Co-instructed "Hacking Thingz Powered By Machine Learning".
May 2016 – Feb 2018
Intel Corporation
Senior Security Engineer
Bengaluru, India
  • Integrated security across the product lifecycle for mobile, IoT, and wearable devices in Intel's New Devices & Wearables Group.
  • Researched embedded hardware and firmware vulnerabilities through fuzzing; presented Droid-FF at Black Hat USA 2016, HITB Amsterdam 2016, and DEF CON 24.
  • Taught adversarial ML security through a TROOPERS17 full-day training, HITB Amsterdam Lab, DEF CON 25 workshop, and BruCon 0x09 workshop.
Oct 2014 – Apr 2016
Citrix
Security Engineer
Bengaluru, India
  • Performed security analysis of products, advised engineers, reviewed bug fixes, and researched emerging vulnerabilities.
  • Researched Android, Mobile Device Management, and Mobile Application Management security.
May 2014 – Oct 2014
EY
Information Security Specialist
Bengaluru, India
  • Conducted mobile OS reviews, web/thick-client/mobile penetration tests, and developer security training.

Conferences

c0c0n
Review Board
Technical review board member
DeFi Security Summit
Review Board / 2025
Technical review board member. Moderated the "AI-Powered Auditing: Hype, Reality, and the Future of Web3 Security" panel at DSS #4, Buenos Aires, November 2025.
BSidesSF
2026
Presented "Saving Bug Bounties from AI Slop"
DEF CON
24 / 25 / 26 / 29 / 34
Speaker, workshop instructor, and AI Village founding organizer. DEF CON 34 Adversary Village lab: "Breaching the Frontier: A Hands-On Lab on AI-Native Vulnerabilities", with the ADV26 tooling release.
Black Hat USA
2016 / 2017
Arsenal presenter for Droid-FF and adversarial ML tooling
TROOPERS
2017
Full-day adversarial ML security training
HITB
Amsterdam 2016 / 2017
Droid-FF talk and adversarial ML lab
BruCon
0x09
Adversarial ML workshop instructor
Shakacon X
2018
Presented "AI Toolkit for Hackers"
Hacktivity
2020
Presented FuzzCube distributed fuzzing platform
Nullcon / Hack In Paris / Hack.lu / PHDays
2016 – 2018
International security talks and workshops

Publications

Podcasts & Talks

Projects

AI red-team infrastructure
ADV26
Blueprint for running open-source reasoning models as agents against an authorized lab range at scale: a GOAD Active Directory range, vLLM-served models, constrained agents over Tailscale, and an operations console with replayable event traces. Released at DEF CON 34 Adversary Village.
View on GitHub →
Android instrumentation
frida-android-hooks
Frida-based runtime instrumentation framework for Android method hooking; widely adopted for dynamic analysis with roughly 400 GitHub stars.
View on GitHub →
Distributed fuzzing
FuzzCube
Kubernetes-based fuzzing infrastructure for scalable, parallel adversarial-input generation and automated red-teaming workflows.
View on GitHub →
Android fuzzing
Droid-FF
Android file-fuzzing framework for automated vulnerability discovery, presented at Black Hat USA, HITB Amsterdam, and DEF CON 24.
View on GitHub →
LLM agents
Kube SRE GYM
An RL environment for training LLMs to diagnose and fix Kubernetes incidents against a real k3s cluster using command feedback.
View on GitHub →
ZK bug bounties
Zero Trust Bounties
Cryptographically verifiable bug-bounty protocol using zero-knowledge proofs; featured in Forbes in September 2025.
View on GitHub →
AI + formal methods
Forge Proof
AI-powered security analysis with formal verification, pairing frontier LLMs with Halmos symbolic execution to find bugs with proof.
View on GitHub →

Awards & Achievements

1st Prize, Team "Stack Too Deep"
Wonderland CTF
November 2025
1st Prize – Security Track
ETHDenver
March 2024
Ethereum Foundation Prize + Finalist
ETHGlobal NYC
September 2023
Software Quality Award Finalist
Intel SW Directors' Council
September 2017