A

Arkose Labs

Deters bot attacks with risk analytics

Senior Machine Learning Researcher

Full-Time
No salary listed
Senior
Bachelor's, Master's, PhD
Pune, Maharashtra, India
HybridHybrid office model with work-from-home days.
AI & Machine LearningAI Research

About the job

Requirements
  • A Bachelor's, Master's, or PhD in Computer Science, Machine Learning, Artificial Intelligence, or a related field is required.
  • Proficiency in machine learning algorithms, neural networks, and model development is required.
  • Experience with machine learning frameworks and libraries such as TensorFlow, PyTorch, and scikit-learn is required.
  • Strong coding skills in Python and R, and familiarity with languages such as C++, Java, or Scala are required.
  • Ability to perform data cleaning, preprocessing, and exploratory data analysis is required.
  • A strong understanding of statistical methods and their applications is required.
  • Familiarity with cybersecurity concepts, threat models, and common attack vectors is required.
  • Experience with security tools and technologies such as intrusion detection systems, firewalls, and malware analysis tools is required.
  • Understanding of network security, bot security, application security, and incident response is required.
Responsibilities
  • Develop and implement machine learning and deep learning models to detect and mitigate security threats.
  • Develop product features using statistical and machine-learning-based algorithms to identify bot and fraud traffic.
  • Clean and preprocess data, conduct exploratory data analysis, and apply statistical methods to analyze security-related data.
  • Stay current with machine learning and cybersecurity research and advancements, and publish research findings in reputable journals and conferences.
  • Collaborate with cross-functional teams to understand project requirements and contribute to solution development.
  • Collaborate with security analysts, software engineers, and other stakeholders to integrate machine learning solutions into security products.
  • Communicate complex technical concepts effectively to technical and non-technical stakeholders.
Desired Qualifications
  • Experience with bot detection, device fingerprinting, behavioral biometrics, or real-time risk scoring.
  • Experience with low-latency, high-throughput machine-learning serving infrastructure.
  • Familiarity with current machine learning research relevant to fraud and security, including adversarial machine learning, anomaly detection, or graph learning.
  • Experience scaling researcher productivity using Agentic AI.

About the company

A

Arkose Labs

View
cbWebsite

Arkose Labs builds cybersecurity tools to stop bot attacks. Its platform combines real-time risk assessments, machine learning analytics, transparent risk insights, and fast attack responses to detect and mitigate bot activity before it can cause damage. A key product, Arkose MatchKey, is described as the strongest CAPTCHA ever made, designed to be highly defensible, usable, and accessible. Unlike some competitors, Arkose Labs focuses on deterring attacks by raising the cost and effort required for adversaries to succeed, making bot attacks more expensive and less attractive. The company earns revenue by charging clients for access to its platform and services, with pricing typically tied to the level of protection, the size of the client’s digital presence, and contract duration. Overall, the goal is to protect digital assets from malicious bots by making attacks harder and more costly to carry out.

Company Size

201-500

Company Stage

Series C

Total Funding

$99.5M

Headquarters

San Mateo, California

Founded

2016

Get referred to Arkose Labs

See people who can refer or advise you

Simplify Jobs

Simplify's Take

What believers are saying

  • Arkose said agentic traffic doubled quarter-over-quarter and surged 1,100% by August 11, 2026.
  • Fifth Inc. 5000 recognition and 257 employees signal growth without staffing collapse.
  • Customer names like Meta, Adobe, Expedia, and Anthropic validate enterprise demand in 2026.

What critics are saying

  • Cloudflare blocked AI training crawlers by default on July 1, 2026, commoditizing traffic controls.
  • DataDome and HUMAN shipped agent-trust platforms in 2026, directly attacking Arkose's category lead.
  • If Titan fails against agentic fraud, Meta and Microsoft can replace Arkose with bundled controls.

What makes Arkose Labs unique

  • Arkose Titan launched January 29, 2026, unifying bots, AI agents, and human fraud defense.
  • Agent Trust Manager classifies intent across three populations, not just automation, as of June 16, 2026.
  • Arkose backs Titan with a $1 million warranty, sharpening enterprise accountability versus peers.

Help us improve and share your feedback! Did you find this helpful?

Benefits

  • Health Insurance
  • Dental Insurance
  • Vision Insurance
  • Life Insurance
  • Disability Insurance
  • 401(k) Retirement Plan
  • 401(k) Company Match
  • Unlimited Paid Time Off
  • Flexible Work Hours
  • Paid Vacation
  • Generous nationwide parental leave policy
  • Employee Assistance Program

Growth & Insights and Company News

Headcount

6 month growth

2%

1 year growth

2%

2 year growth

2%
Prosopo
Sep 1st, 2026
Gartner and bot management: the Magic Quadrant that does not exist.

Gartner and bot management: the Magic Quadrant that does not exist. There is no Gartner Magic Quadrant for bot management. There is one for Cloud WAAP, and bot mitigation is a criterion inside it. Where each vendor appears, and how to read the two analyst views together. The Gartner Magic Quadrant for bot management is one of the most requested documents in this category, and it does not exist. The reason it does not exist is useful for understanding how the market is structured. Gartner covers this space in two places, neither of which is a bot management Magic Quadrant. The Magic Quadrant for Cloud Web Application and API Protection is the relevant quadrant. Cloud WAAP is Gartner's category for cloud-delivered services combining a web application firewall, DDoS mitigation, bot management and API protection into one offering. Cloudflare, Akamai, Imperva, F5, AWS, Fastly and Radware are the vendors you find there. Bot mitigation is a criterion inside a broader evaluation. Gartner Peer Insights runs a separate Bot Management review category, built on customer reviews rather than analyst placement. DataDome, HUMAN, Kasada, Arkose Labs, Netacea, CHEQ, F5, Radware, Imperva and Cloudflare all have profiles there. So when a vendor cites Gartner on bot management, it is worth asking which of those two they mean, because they carry very different weight. Put the lists next to each other and the difference is stark. Forrester's Q2 2026 Bot and Agent Trust Management Wave names DataDome, HUMAN and Kasada as Leaders, Arkose Labs, CHEQ and Netacea as Strong Performers, and hCaptcha and Google reCAPTCHA Enterprise as Contenders. Cloudflare and Akamai are absent. Gartner's Cloud WAAP quadrant is built around Cloudflare, Akamai, Imperva, F5 and AWS. The Forrester specialists are mostly absent. Neither list is incomplete. They are answers to different questions. Gartner's question is roughly: which vendor can consolidate my web application security into one cloud service? That question structurally favours the CDN and WAF platforms, because consolidation is what they sell. Forrester's question is: which vendor is best at establishing trust for bots and agents specifically? That question structurally favours the specialists, because depth is what they sell. If your procurement process picks one analyst view and treats it as the market, you get a predictable failure. Use only the WAAP quadrant and you buy a bundled product that handles commodity automation well and targeted attacks poorly. Use only the Wave and you may buy a specialist to run alongside a WAF you already pay for, duplicating what you have. A practical order of work, drawn from watching these evaluations run: Start with what is already in the contract. If you have Cloudflare Enterprise or an Akamai CDN, you already own bot mitigation. Turn it on properly and measure it before you buy anything. Sometimes that is the whole project. Then find out what gets through. The bundled layer handles the commodity traffic. What survives is the question. If what survives is stealth-headless frameworks on residential proxies, solver services, or real-device farms, no amount of WAAP consolidation fixes it, and the Wave is your list. Then decide whether you want to operate it. DataDome, HUMAN, Kasada and Arkose all include a managed service. That is a product decision rather than a pricing one, and it determines who authors the rule during a live incident. Then look at Peer Insights, not the placement. Vendor placement tells you about vision and revenue. Reviews tell you what daily operation is like. Forrester's own Wave picks up the same themes that appear in reviews: CHEQ's numeric reason codes, Kasada's limited UI customisation, hCaptcha's clunky management interface. One more thing worth knowing if you are searching for research. Forrester renamed this market in late 2025. What was bot management, bot mitigation and bot detection is now bot and agent trust management, on the grounds that AI agents fall between humans and bots and a binary block-or-allow decision loses real customers. Gartner has not adopted that rename as of Q3 2026. Gartner still says bot management. So the same market currently has two analyst names, and searching for one will not find research filed under the other. If you are doing a literature review before a purchase, search both. Prosopo is not currently in either evaluation. The Wave applies a standalone product revenue threshold, and a Gartner Peer Insights presence is on its roadmap. What Prosopo compete on today is the specific capability both analyst houses repeatedly mark vendors down for. Forrester criticised CHEQ for numeric reason codes that are difficult to interpret and reCAPTCHA Enterprise for lack of detail and limited drill-down. Arkose customers asked for more self-service configuration. Kasada customers reported limited UI customisation. Every Prosopo block comes with a reason you can read, not a numeric score or a policy label. Each one is a string your team can grep, like Solver service detected or Inconsistent hardware readings for device, and you can look up the underlying event by request ID and export it to whatever SIEM you already run. That is a design choice rather than an analyst rating, and you should weigh it accordingly.

VocoLife
Aug 11th, 2026
Arkose Labs: Inc. 5000 for 5th year as agentic AI surges.

Arkose Labs: Inc. 5000 for 5th year as agentic AI surges. 1h ago · 0:00 listen · Source: citybiz Summary. Arkose Labs has been named to the Inc. 5000 for the fifth time. What's interesting is this cybersecurity company is expanding its technology as agentic AI traffic surges. Arkose reports an 1,100% increase in AI agent traffic across its network from the first quarter to the second quarter. This creates new security challenges as businesses use autonomous software for tasks and transactions. The company is addressing this with Agent Trust Manager, a new part of its Titan security platform. This technology determines if automated traffic is a customer, an authorized AI agent, or an adversary. It then decides whether to trust, challenge, or block the activity. The system builds on existing device intelligence and behavioral biometrics. The bottom line is this helps businesses accommodate legitimate AI activity without giving fraudulent agents access. This is an AI-generated audio summary. Always check the original source for complete reporting.

Arkose Labs
Aug 11th, 2026
Arkose Labs named to Inc. 5000 List of America's fastest-growing private companies for fifth year.

Arkose Labs named to Inc. 5000 List of America's fastest-growing private companies for fifth year. As Agentic AI Traffic Surges 1,100% Quarter over Quarter, the Company is Helping Companies Embrace Agentic Commerce with New Arkose Agent Trust Manager SAN MATEO, Calif. - August 11, 2026 - Arkose Labs, the leader in agent trust, today announced it has once again been named to Inc. Magazine's prestigious 2026 Inc. 5000 List of America's fastest-growing companies. This marks the company's fifth appearance on the annual ranking, driven by three-year revenue growth. Arkose Labs' customer base includes Anthropic, Microsoft, Meta and Roblox, plus other Fortune 500 leaders in financial services, online dating, travel and gaming. "Arkose Labs' inclusion on the Inc. 5000 list for five years is indicative of our ability to meet the changing threats facing top business-to-consumer companies around the globe," said Frank Teruel, Chief Operating Officer of Arkose Labs. "As agentic AI escalates automated fraud at a scale and sophistication never seen before, we continue to help our customers meet the challenge and prepare for future threats." The company's founder and CEO Kevin Gosschalk, a recognized expert on agentic AI, recently published a book on its growing threat called After Bots. Built on 10 years of hands-on experience, After Bots is the definitive playbook for what comes next in trusting and controlling AI agents. It provides actionable guidance on how to welcome authorized agents and the revenue they represent while shutting out the malicious ones. "Arkose Labs has seen agentic traffic increase 1,100% from Q1 to Q2, and we expect agentic commerce to really skyrocket during the coming holiday season," said Kevin Gosschalk, founder and CEO of Arkose Labs. "Arkose Agent Trust Manager will help our customers truly embrace this opportunity while keeping malicious activity at bay." A critical addition to its Arkose Titan(TM) platform, Arkose Agent Trust Manager recognizes whether agentic traffic represents a customer, an authorized agent or an adversary to determine whether they should be trusted, challenged or stopped. Arkose Agent Trust Manager runs on top of the device intelligence, behavioral biometrics and adaptive challenge telemetry already embedded in Arkose Titan. Arkose Labs' overall customer base includes: * 2 of the top 4 traditional banks by assets, securing the financial transactions of 250+ million end users worldwide * 2 of the 4 largest social media platforms, including Meta, serving 4+ billion accounts across the globe * 4 of the top 10 dating apps globally, ranked by downloads and revenue * 2 of the top 4 travel sites, including Expedia, helping to protect $200+ billion in gross bookings * 3 of the top 10 gaming companies, including Roblox, with 250+ million active players around the globe For the full Inc. 5000 list, honoree company profiles, and a searchable database by industry and location, visit: www.inc.com/inc5000. About Arkose Labs. Arkose Labs is the global leader in agent trust and control and helps companies identify whether a site visitor is a customer, AI agent or adversary. Arkose Titan, its unified platform, provides controls over how they are trusted, challenged or stopped accordingly. In an era when AI agents are changing the economics of online fraud, Arkose Labs protects against persistent, massive scale AI-fueled abuse rather than just isolated, bot-driven attacks. Trusted for over 10 years by global financial services, technology, and travel companies including Anthropic, Meta, Roblox and Expedia, the company has been named to the Deloitte Technology Fast 500 for five consecutive years. Arkose Labs backs the Arkose Titan platform with the industry's only $1 million warranty program. The company is headquartered in San Mateo, California, and has research teams around the globe. Media Contact:

Associated Press
Jun 16th, 2026
Arkose Labs launches Agent Trust Manager to classify AI traffic as customer, authorised agent or adversary

Arkose Labs has launched Arkose Agent Trust Manager, a new tool that classifies AI-driven traffic as customers, authorised agents or adversaries. The company reports agentic traffic is doubling quarter-over-quarter among its customer base, which includes Meta, Adobe and Expedia. Agent Trust Manager operates on a three-population framework, using device intelligence, behavioural biometrics and adaptive challenge telemetry to classify sessions. It then enforces responses through a five-step spectrum including allow, monitor, challenge, throttle and block. The system addresses a critical gap as 74% of adversarial sessions spoof browser values and 69% fake macOS environments. The tool runs on Arkose Labs' existing Arkose Titan platform, which the San Mateo-based company backs with a $1 million warranty programme.

Arkose Labs
Jun 16th, 2026
Arkose Agent Trust Manager helps companies accelerate revenue by classifying and controlling agentic traffic as a customer, an authorized agent or an adversary.

Arkose Agent Trust Manager helps companies accelerate revenue by classifying and controlling agentic traffic as a customer, an authorized agent or an adversary. LAS VEGAS, NV (Identiverse) - June 16, 2026 -Arkose Labs, the leader in agent trust, today announced the launch of Arkose Agent Trust Manager. It arrives just as enterprises face an unprecedented surge in AI-driven traffic that existing bot and fraud defenses cannot handle. A critical addition to its Arkose Titan(TM) platform, Agent Trust Manager recognizes whether agentic traffic represents a customer, an authorized agent or an adversary to determine whether they should be trusted, challenged or stopped. Companies including Meta, Adobe and Expedia already work with Arkose Labs to protect high-value customer journeys against both traditional and agentic threats. It prevents attackers from slipping through while ensuring both human and agentic customers can quickly and easily transact online. "Among our customer base, we are already seeing agentic traffic double quarter-over-quarter," said Kevin Gosschalk, founder and CEO of Arkose Labs, "The companies that successfully combat them won't be the ones that simply define an agent control policy. It will take real time, automatic enforcement of those policies without a human ever having to get involved. This enforcement gap is exactly what Arkose Agent Trust Manager closes." Agent Trust Manager runs on top of the device intelligence, behavioral biometrics and adaptive challenge telemetry already embedded in Arkose Titan. This telemetry reveals the scale of the impersonation problem: 74% of adversarial sessions spoof browser values and 69% fake macOS environments, making behavioral and device-level analysis essential to accurate classification. Arkose Labs' approach is that intent is not inferred from identity, but determined by behavior. Agent Trust Manager provides the trust classification, enforcement and control layer that runs on top of Arkose Titan's existing threat signal stack, including ACTIR threat intelligence. It is positioned where legitimate users, authorized business agents, ambiguous scrapers and malicious adversaries all arrive at the same flow. Not bot or human. Three distinct populations, each requiring a different response. Agent Trust Manager operates on a three-population framework that reflects the reality of today's traffic composition. * Self-Disclosing Good Agents: Verified crawlers, partner integrations and cooperating platforms that publish IP ranges and sign HTTP headers. Identified through disclosure standards and cryptographic verification. * Non-Disclosing Good Agents: Legitimate AI assistants including Claude computer use, ChatGPT agent mode and agentic browsers that operate on behalf of real users but offer no self-disclosure. * Malicious Adversaries: Account takeover campaigns, fake account factories, payment fraud operations and coordinated swarms. Cloud-hosted at scale, almost always impersonating legitimate browser environments via residential proxies Agent Trust Manager focuses on classification, enforcement and control. Agent Trust Manager relies on a combination of device intelligence, behavioral biometrics, and adaptive challenge telemetry to determine how traffic should be classified and enforced. The complete loop preserves the experience for legitimate users. * Classify: Multi-signal evaluation resolves every session into one of four classes including human, self-disclosing good agent, non-disclosing good agent, or malicious adversary. The intent detection runs on top. It features an array of agentic AI detections to cover the full spectrum of agent traffic across five detection layers. * Enforce: A five-step enforcement spectrum including Allow, Monitor, Challenge, Throttle and Block that matches each session to the proportional response, minimizing friction for legitimate users while raising cost for adversaries. Proof of Work and AI-resistant challenges apply economic deterrence on the same surface where classification occurs. * Control Customer-defined policy by endpoint, by population and by use case. Continuous intent integrity monitoring re-classifies sessions when behavior drifts from declared scope, automatically escalating enforcement when an agent authorized for one activity begins probing restricted flows. Unlike fragmented point solutions, Arkose Titan provides defense-in-depth through intelligent detection and adaptive mitigation against both traditional and emerging AI threats, including agentic AI. By defending a company's entire digital experience and customer journey, Arkose Titan makes attacks economically unsustainable for perpetrators. It helps companies build their businesses uninterrupted and avoid fraudulent account chargebacks and customer reacquisition costs. About Arkose Labs. Arkose Labs is the global leader in agent trust and control and helps companies identify whether a site visitor is a customer, AI agent or adversary. Arkose Titan, its unified platform, provides controls over how they are trusted, challenged or stopped accordingly. In an era when AI agents are changing the economics of online fraud, Arkose Labs protects against persistent, massive scale AI-fueled abuse rather than just isolated, bot-driven attacks. Trusted for over 10 years by global financial services, technology, and travel companies including Anthropic, Meta, Roblox and Expedia, the company has been named to the Deloitte Technology Fast 500 for five consecutive years. Arkose Labs backs the Arkose Titan platform with the industry's only $1 million warranty program. The company is headquartered in San Mateo, California, and has research teams around the globe.