← Back to home

Company

We built the security layer that browser proctoring forgot.

The certification Akshay spent months earning was being sold by proxy rings for $200, pay after pass. At Cornell Tech's Startup Studio, the founding team found each other: four people who had all hit the same wall from different angles.

2025
Founded at Cornell Tech
30+
Attack vectors blocked
$300K+
Bug bounties, Akshay
Alpha
Live on production

Origin

The credential that couldn't be trusted.

In 2022, Akshay built upskillfinder.com around one hypothesis: stacking top-tier cybersecurity certifications was the cleanest signal of competence the industry had. By 2024, he had 30 of them: OSCP, CEH, the full stack. The plan was to build an authority profile so strong that no recruiter, client, or enterprise customer could dismiss it.

Then he found the proxy-ring ads. The same credentials, the ones he'd spent months earning, were available to order from syndicated proxy networks. Pay $200 after the test clears. The proxy sits the exam remotely. The “candidate” gets coffee. Full credential delivered. Zero skill transferred.

The certification had become a liability. His hard-earned OSCP was indistinguishable from one bought for $200. Every signal he'd built was now equally accessible to someone who had never opened a terminal.

At Cornell Tech's Startup Studio in November 2025, Akshay found his co-founders. Sudhakar had spent eight years building secure cloud infrastructure and had watched assessments fail at scale. Divya had spent his career red-teaming AI systems, including the exact overlay architectures now being sold as cheating tools. Sanjay had encountered the proxy-ring economy first-hand while preparing for a graduate admissions exam.

The common thread wasn't the specific failure mode. It was the layer at which every existing solution tried to stop it: the browser, the webcam, the application. None of it was close to where the cheating actually happened.

They built Aiseptor at the network layer instead.

Founders

Built by people who broke the thing they're now defending against.

Offensive security, AI red-teaming, secure infrastructure, and enterprise GTM: each domain is load-bearing for the network-layer thesis to work.

Akshay Aggarwal

Akshay Aggarwal

CEO & Founder

LinkedIn ↗

Akshay built upskillfinder.com to prove a hypothesis: that accumulating every top cybersecurity certification was the cleanest credibility signal in the field. He earned 30+. Then he found the proxy rings selling the same credentials for $200 each, pay after pass, and built Aiseptor to make credentials mean something again.

Ten-plus years in offensive security. Over $300,000 in bug bounties across Fortune-500 programs. Cornell Tech master's program. Leads Aiseptor's technical architecture and threat model.

Upskillfinder.comCornell Tech30+ Certifications$300K+ Bug Bounties
Sudhakar Padmanaban

Sudhakar Padmanaban

Chief Product Officer

LinkedIn ↗

Eight-plus years architecting backend systems and secure cloud infrastructure across cybersecurity, networking, and cloud-native platforms. Saw assessments fail at scale from the inside, in enterprise environments where the security tooling was well-funded but the assessment layer was an afterthought.

Johnson Cornell Tech MBA (Class of 2026), where he met the founding team. Owns Aiseptor's product surface, from the enterprise dashboard and proctor console down to the agent lifecycle on Windows and macOS.

Cornell Johnson Tech MBA8+ yrs InfrastructureCloud SecurityDistributed Systems
Divya Bhanushali

Divya Bhanushali

Chief AI Officer

LinkedIn ↗

AI Safety and Red-Teaming Fellow at Cornell Tech. Spent his career reverse-engineering the attack surface of AI systems, including the overlay architectures now productized as the leading cheating tools, before they reached public distribution.

Selected for the inaugural cohort of Cornell's SETS AI Red Team Clinic, where his team stress-tested the NYC Health Department's AI tools and built a semi-automated prompt-testing pipeline that significantly expanded the team's assessment capacity. Previously software engineer at ZS Associates and intern at Intuit. Leads Aiseptor's adversarial-AI research.

Cornell TechAI Safety & Red-TeamingZS AssociatesIntuit
Sanjay Ram

Sanjay Ram

Chief Business Officer

LinkedIn ↗

Encountered the proxy-ring economy first-hand while preparing for a high-stakes standardized exam. Saw the ads, pulled the thread, and realized the problem was structural: any credential worth holding was worth gaming.

Five-plus years across AI, cloud, and enterprise go-to-market. Cornell Tech Startup Studio. Runs Aiseptor's partnerships, platform embedding deals, and enterprise relationships, including 3 signed pilots with assessment platforms.

Cornell TechEnterprise GTMAI & CloudPartnership Strategy

We met in Cornell Tech's Startup Studio in November 2025. The company was incorporated before the semester ended.

Advisors

Depth in the two domains the thesis depends on.

The global assessment industry and offensive cybersecurity. Three advisors, each with decades inside the system Aiseptor is rebuilding.

Graham Hudson

Graham Hudson

Advisor, Assessment

LinkedIn ↗

Chair of the e-Assessment Association (eAA), the global membership body for assessment technology. Thirty-five years at the intersection of technology and high-stakes qualifications, including leading a £3.5M cross-regulator R&D programme across the four UK exam regulators and overseeing the first live digital marking of UK national examinations.

Former senior roles at the UK's Qualifications and Curriculum Authority (QCA) and DRS Data Services. Brings the regulatory and institutional gravity that enterprise assessment buyers require.

Chair, e-Assessment AssociationGA Partnership LtdQCA, UK Gov35+ yrs Assessment
Sunny Nehra

Sunny Nehra

Advisor & Board, Offensive Security

LinkedIn ↗

Founder and CEO of Secure Your Hacks. Widely recognized as India's leading ethical hacker, having identified critical vulnerabilities in the Indian Army, RailTel, Kotak Securities, and PSPCL, among others.

Senior cybersecurity consultant to India's Ministry of Home Affairs. Expert witness in cybersecurity court proceedings. Trains law enforcement and judicial officials in digital forensics and cyber law. Anchors the offensive-security depth that sets Aiseptor's detection ceiling.

Founder, Secure Your HacksMHA Cybersecurity ConsultantIndian Army / NavyIndia #1 Ethical Hacker
Sangam Singh

Sangam Singh

Technical Advisor

LinkedIn ↗

Founder and CEO of Infinite Potential. Delivered over $20M in technical projects for the Indian Army, Navy, and Air Force, in environments where security failures are not post-mortems but operational consequences.

Newcastle University graduate. Brings large-scale sovereign-grade deployment, operational resilience, and defence-sector security experience to Aiseptor's roadmap and go-to-market in regulated markets.

Founder & CEO, Infinite PotentialIndian Army / Navy / Air ForceNewcastle University$20M+ Gov Deployments

Traction

Where we are, no spin.

Alpha. Users on production, no revenue yet, bootstrapped roughly $20K to date. The signal that matters is what we've built, who's validated it, and who's signed the LOI.

100+

C-suite assessment and talent leaders interviewed before a single enterprise feature shipped.

3 pilots

Three signed paid pilots with assessment platforms. Active deployment, integration scope locked.

2 pilots

Two live pilot runs, 100+ candidates each, at an Ivy League research university.

30+

Distinct AI-cheating and exfiltration attack vectors reverse-engineered and neutralized in the MVP.

Patent pending

Ephemeral user-space network enclave, deployable in under 60 seconds with no kernel drivers and no persistent install.

Alpha live

On production, Windows and macOS. Per-session transparent pricing. Pre-revenue by design.

Why now

Three forces converging.

The category is being recreated. We intend to author the primitive it runs on.

01

The cheating tool moved below the browser.

Invisible overlays and on-device LLMs reached mainstream distribution between 2023 and 2025. Every proctoring product architected before that window is solving the wrong threat model. Retrofitting won't close the gap; the tool now lives below the browser, below the OS install registry, and entirely outside the webcam's field of view.

02

The required skill stack is absent from EdTech and HRtech.

Building a user-space security enclave that deploys in 30 seconds on a candidate's personal laptop, with no driver, no admin permission, and no persistent footprint, requires deep expertise across cryptographic tunneling, real-time access policy enforcement, OS-level process introspection, and hardware telemetry. That combination is not a product roadmap item for any existing assessment company.

03

Enterprise security giants can't build disposable BYOD tools.

Palo Alto, CrowdStrike, and Zscaler architect for managed fleets and 30-day deployments. A candidate's personal MacBook for a 90-minute exam breaks every assumption in their model. The innovator's dilemma leaves the seam open, and that seam is exactly where AI cheating lives.

Press

Press kit & media.

Journalists covering the AI-cheating beat are welcome to request a technical briefing with the founding team. We can arrange demos, provide threat-model context, and connect with the advisors.

Coverage

No press mentions yet; we're early. When coverage lands, it will appear here and in the press kit. The 2026 AI cheating research report is the best starting point for understanding the threat landscape and why the network-layer approach is technically necessary.

Building the security layer for assessments that actually matter.

Platform partnerships, enterprise pilots, press inquiries, and hiring conversations all route to the founding team. No sales queue.