GRC Engineer
About the job
- Have 5+ years of experience in security, compliance, automation, or engineering functions in a fast-paced environment.
- Have experience designing or implementing AI-powered automation and agentic workflows, including understanding the risks of non-deterministic systems and how to govern them.
- Have deep experience implementing Infrastructure as Code using Terraform or Pulumi and Policy as Code using Open Policy Agent or Rego.
- Be proficient in Python, Go, or other scripting languages for automation, API interactions, and data parsing.
- Be comfortable working with REST APIs to integrate disparate security tools.
- Have proven experience building modern web applications using JavaScript or TypeScript and React.
- Have access to information protected under U.S. export control laws only with authorization to receive controlled software or technology without sponsorship for an export license.
- Develop and implement automated solutions to improve GRC processes and operations, integrating with existing security, engineering, and artificial intelligence tools.
- Build modules and maintain the GRC platform and related initiatives.
- Architect and maintain a technical governance framework for the safe deployment of autonomous artificial intelligence agents.
- Ensure agentic workflows operate within deterministic action gates and predefined risk thresholds.
- Support Cloudflare’s security assessments, including SOC 2, ISO 27001, PCI DSS, and FedRAMP.
- Maintain audit readiness through automated drift detection and self-healing configurations across the global edge network.
- Work cross-functionally with Legal, People, Engineering, and Finance teams to integrate security into company operations.
- Integrate security into a paved-road model so compliance is the default state for every developer.
- Demonstrated passion for security and software development through personal projects, open-source contributions, or active participation in the security research community.
- Experience building with Cloudflare’s developer platform and products such as Cloudflare Workers, R2, D1, or Workers AI.
- Effective collaboration across frontend and backend boundaries.
About the company
Cloudflare operates a network that protects websites, delivers online content and runs applications closer to their users. Businesses connect websites and internal systems to its services to filter malicious traffic, control access and improve performance, with security and computing running on shared infrastructure. Developers combine application code, object storage and databases with the network used for content delivery and traffic protection. Cloudflare earns revenue from subscriptions and usage charges, with free plans for basic features and negotiated contracts for larger organizations. Cloudflare was founded in 2009 and is headquartered in San Francisco, California.
Company Size
1,001-5,000
Company Stage
IPO
Headquarters
San Francisco, California
Founded
2009
Get referred to Cloudflare
See people who can refer or advise you
Simplify's Take
What believers are saying
- Demand is growing: Cloudflare reported $696 million revenue, up 36%, for the June 2026 quarter.
- New parents get paid bonding time: qualifying Cloudflare employees receive up to sixteen paid weeks globally.
- Attack protection has predictable billing: Cloudflare provides unmetered DDoS mitigation across all plans, including its free tier.
What critics are saying
- Employees faced a major workforce reduction: Cloudflare announced more than 1,100 global job cuts in May 2026.
- One software fault spread widely: Cloudflare's bot-management configuration broke core network traffic in November 2025.
- Free users still incur costs: most Cloudflare customers pay nothing, while the company bears incremental network expenses.
What makes Cloudflare unique
- Services share network locations: Cloudflare runs its security and delivery stack in every data center.
- R2 makes outbound bandwidth free: Cloudflare charges for storage and requests, without fees for direct Internet data transfers.
- Workers handles deployment and scaling: developers publish apps across Cloudflare's global network without managing servers.
Help us improve and share your feedback! Did you find this helpful?
Benefits
- Competitive salaries
- Take-what-you-need paid vacation policy
- Comprehensive health plans and benefits
- Paid maternity and paternity leave
- Commuter and ride share options
- Returnships
Growth & Insights and Company News
6 month growth
0%1 year growth
0%2 year growth
0%DNS filtering can help protect children online, but parents need more than a filtered connection. 3 minutes read Families looking for ways to reduce children's exposure to harmful websites can start with controls built into their internet connection. DNS filtering can prevent a device from reaching certain domains, including known malware sites and, depending on the service, websites classified as containing adult content. It offers a useful layer of protection, although it cannot determine everything a child will encounter online. The subject came up during the Safaricom Cybersecurity Summit 2026, where Cloudflare's Graham Turnbull recommended the company's family DNS service as one option for households. Cloudflare provides the resolver addresses 1.1.1.3 and 1.0.0.3 for filtering known malware and adult-content domains. The service is a basic network-level control, not a complete parental-control system. What DNS filtering does. The Domain Name System, or DNS, helps devices find the network address associated with a website's domain name. When someone enters a web address, the device typically makes a DNS request to find where that domain should resolve. A filtering resolver checks the request against its categories or lists and can block domains that match the selected rules. Cloudflare's family resolver is designed to block domains associated with malware and adult content. A household can configure compatible network equipment to use the resolver, allowing supported devices on that network to receive the filtering without having to configure each browser separately. The exact setup depends on the router, operating system and network configuration. How family filtering services work. Using a family DNS resolver can be one of the simpler ways to add a baseline filter to a home network. Parents should follow the service provider's configuration instructions and check that devices are actually using the intended resolver. A device may use a different DNS setting when it connects to another network or when software changes how its internet requests are handled. The same approach can be useful for reducing exposure to known malicious domains. Blocking a domain at the DNS layer can prevent a connection before a user reaches the site, but it depends on the service's classifications and the domain being recognised. New, altered or incorrectly categorised domains may not be handled as a family expects. What DNS filtering cannot prevent. DNS filtering does not inspect every page, image, message or interaction a child encounters. It may not block harmful material hosted on an otherwise legitimate website, and it cannot by itself prevent cyberbullying, manipulative conversations, inappropriate messages or the sharing of personal information. Nor should parents assume that every dangerous link will be recognised by a filter. It is also not a substitute for device security or account controls. Children may use mobile data, another Wi-Fi network or applications that apply their own network settings. Parents should understand how the devices in their household connect to the internet and avoid treating one network setting as a guarantee. How to build a broader online-safety plan. Families can combine DNS filtering with age-appropriate device settings, app-store restrictions, browser safeguards and controls offered by the platforms children use. Keeping devices and applications updated can also reduce exposure to known security weaknesses. The appropriate combination depends on the child's age, the devices available and the kinds of online activities involved. Technical controls work best when children understand why they exist. Parents can discuss suspicious links, requests for personal information, unwanted contact and what to do when content makes them uncomfortable. They should also make it easy for children to ask for help without fearing an automatic punishment for reporting a problem. DNS filtering can reduce exposure to selected categories of harmful websites with relatively little ongoing effort. Its value comes from treating it as one component of a broader approach that combines technical settings, responsible supervision and open communication. Real ESG impact doesn't happen in panels alone, it happens in the rooms where financiers, operators, and policymakers actually align. Our GreenShift Forum 2026 cuts the noise, bringing together the people rewiring Africa's sustainability and energy frameworks for one focused day in Nairobi. Secure your seat. Go to TECHTRENDSKE.co.ke for more tech and business news from the African continent and across the world. Facebook Comments
Rebuilding companies for the agentic era. Oct 13, 2026, 1:12pm PDT Companies have spent the past few years figuring out how to put AI to work. In 2026, some are confronting a bigger question: What should a company look like when AI can understand context, coordinate work and act across systems? The Information's Co-Executive Editor Amir Efrati sat down with leaders from three companies at the forefront of this shift to discuss how agentic AI is changing how companies work and operate. * Owen Jennings, Business Lead & Executive Officer, Block * Stephanie Cohen, Chief Strategy Officer, Cloudflare * Surabhi Gupta, CTO, Klaviyo Redesigning the operating model Block and Cloudflare offered two ways to rethink the operating model. For Cloudflare, Stephanie Cohen said the company had to rethink the work itself, not simply add AI tools onto existing processes. "If we didn't redesign how work was done, we weren't going to actually change what went on at Cloudflare," she said. "The only way for us to redesign the factory floor was to get rid of whole functions." At Block, the answer was to rebuild its functional model so engineering, product, design and sales teams work across Square, Cash App and Afterpay on shared capabilities. Owen Jennings said the change came about when AI agents became capable of writing production code in complex systems, prompting the company to ask how work should move when agents can share company context with humans. "We said, based on where the tools are now and based on where we think it's going, what functions do we need at the company? How do we want to organize?" recalled Jennings. The shift is showing up in customer experiences, with Square and Cash App surfacing tools, insights or actions at the right moment, rather than relying on static, one-size-fits-all surfaces. Block shipped about three times as many Square features in the first half of 2026 as a year earlier, he said. "There's a lot of conversation around automating things for the sake of it," said Jennings. "For us, it's about building useful stuff for customers." Humans still matter Companies leaning hardest into agents aren't necessarily trying to replace humans. Cloudflare hired 1,111 interns this year, including AI-native talent who Cohen said worked on "everything," including old-fashioned automation no one had made time to build. "By teaching them the company and allowing them to have access to our system, they've shown us a million ways that what we were doing didn't make sense," she said. Block has also built a workspace for working alongside agents. Buzz, an open source tool the company has released for people and AI agents, gives each agent its own identity, permissions and context. "The idea is not to remove people from the loop, but to make agent work visible and steerable so humans can focus on judgment, taste and customer understanding," Jennings said. Klaviyo wants engineers closer to customer outcomes rather than assigned tickets. "One of the things we've found very effective is having the engineers sit side-by-side with the partner functions, and have them become domain experts," said Klaviyo's Surabhi Gupta. Cloudflare runs hands-on workshops that are deliberately low-pressure, and names an AI "champion" on each team, typically someone already experimenting. "It's easier to ask a peer than to ask someone else who you feel like is not gonna understand why you don't know how to do everything," Cohen said. Companies also need infrastructure connecting agents to systems, data and context. Cohen described Cloudflare's approach as a "harness": a layer connecting AI models with security, systems of record, company context and employee identities, while letting the company switch models as needed. "Part of the problem is people started with the hammer and were like, 'here's this hammer, build this house.' The reality is you have to design the house," said Cohen. The new metrics Agentic AI is also creating new metrics to measure success. At Block, development velocity - the pace that new features are shipped - has become a key metric. Gupta says Klaviyo measures this too, as well as cost and hours saved as a result of agents. One metric no one on the panel could predict: headcount, though Cohen felt optimistic about growth. "My guess is we'll have more employees [in five years]," she said. "The human world is human-shaped, and humans want to interact with more humans."
Making AI pay its way, with Cloudflare. Tuesday, October 13th, 2026 - 1:00 am A podcast interview with Lara Cohen VP, Strategic Partnerships, Media, Creators & AI For every human reader they send back to publishers, AI bots scrape thousands of pages, according to Cloudflare. That's not what publishers agreed to when they started letting search engines crawl their sites in exchange for traffic, says Lara Cohen, VP of strategic partnerships for media, creators and AI at Cloudflare, on this week's episode of AdExchanger Talks. Cloudflare, which routes roughly 20% of global internet traffic, is trying to help publishers fight back - and it's an unlikely hero. As a network security company, Cloudflare is better known for providing CDN services, fending off hackers and helping with domain management. But millions of publishers sit on Cloudflare's network, and it could see their human traffic drying up. Cloudflare isn't anti-AI, Cohen says. Far from it. But "in order for AI to thrive," she says, "original human content also has to thrive." And so Cloudflare has built dashboards that show publishers how often large language models like ChatGPT and Claude pull their content, which URLs are popular and how many of their visitors are human. Publishers can then choose to block the bots - or get paid. Last month, Cloudflare launched a beta of "pay per use," a new feature that takes the place of "pay per crawl," its earlier plan to charge AI companies every time a bot scraped a page. The thinking now is that LLMs should pay for the value they actually get from a publisher's content, say by using it to answer a user's question. AI companies make an offer, and publishers decide whether to take it. Buyers have to play fair, though. "They have to declare their crawler," Cohen says. "They have to be showing up honestly." But don't expect AI money to replace ad revenue dollar for dollar. "I can't turn back the hands of time and make it 1995," Cohen says. "What we're looking for in the marketplace is good actors on both sides who are truly trying to experiment and get this right." Also in this episode: Dealing with the "Google problem," Cohen's theory on how local news has a chance to thrive in the AI era and why she'll never let an AI agent go shoe shopping for her. Tagged in:
Peec AI named fastest-growing AEO vendor by Ramp; expands North American footprint with New York office. Data from Ramp Rate reveals Peec AI quadrupled market adoption to 19% while accelerating ARR to $15M, with the second-highest adoption among enterprise companies (21%) and strong traction with growing marketing teams and agencies. Berlin, BERLIN, Oct. 13, 2026 (GLOBE NEWSWIRE) - Peec AI, a leading all-in-one AI search visibility platform for Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO), today announced it has been recognized as the fastest-growing vendor in the AEO category according to market spend data published by Ramp Rate. The company's market adoption rate grew from 4% in March 2025 to 19% in September 2026, driven by widespread adoption across enterprise organizations, mid-market businesses, growth marketing teams, and SEO agencies. During the same period, Peec AI accelerated its Annual Recurring Revenue (ARR) from $10M to $15M in just three months. "It's an honor to be recognized by Ramp Rate for our growth over this year. Ramp's data shows Peec AI is a top AEO vendor in enterprise adoption, with a 21% adoption rate among enterprise companies," stated Malte Landwehr, CPO and CMO of Peec AI. Peec AI ranks first in the AEO category in monthly adoption growth, at +0.84% per month. Peec AI also shows strong traction across mid-market (17%) and small-to-medium businesses (20%). Ramp Rate's dataset tracks corporate spend across more than 70,000 businesses. Enterprise-grade AI visibility, for teams of all sizes Throughout 2026, Peec AI expanded its platform with features built to enterprise standards, from brand perception analysis to log file integrations. Most features are available to teams of every size, whether they're on a self-serve plan or part of an enterprise deployment: * Comprehensive LLM ecosystem tracking: Monitors brand mentions at scale across major AI engines, with broad multi-model coverage of ChatGPT, Google AI Overviews, Google AI Mode, Microsoft Copilot, Gemini, Claude, Grok, Qwen, Deepseek, Mistral, Perplexity, and Meta Spark. * Brand perception and fact-checking: Analyzes AI answers to show how engines like ChatGPT, Gemini, AI Mode, and Claude describe a brand. Market, objections, and fact-checking diagnostics surface the attributes AI answers associate with a brand, identify the buying objections AI platforms repeat, and cross-reference AI-generated claims against company-supplied facts. * Actionable visibility KPIs: Peec AI provides granular reporting across core AI performance metrics, measured on every tracked AI answer, including visibility, sentiment, multi-model citation tracking, share of voice (SoV), and AI referrals. * Developer and agency integration: Provides programmatic API access to AI visibility data and a Model Context Protocol (MCP) integration for AI agent workflows like Claude Code, paired with customizable visualization tools like Power BI and Looker Studio. * Analytics and log file integrations: Peec AI integrates with Google Analytics and CDN log providers like Vercel, Cloudflare, AWS, and WordPress, connecting AI visibility to AI referral traffic and AI crawler activity. * Predictable scaling: Enterprise tiers include unlimited seats, SSO integration, multi-brand management, and multi-language tracking for international operations. On top of that, enterprise customers of Peec AI get a dedicated account manager and a named Customer Success Manager (CSM) alongside custom onboarding and prioritized support channels. Custom plans (starting around $499+/month) include direct phone support and named personnel to guide multi-market tracking architectures. Expanding footprint in North America Peec AI has opened a flagship U.S. office in New York City to support its expanding enterprise user base across North America. The hub will serve North American clients, agency partners, and strategic enterprise accounts. About Peec AI Peec AI helps marketing teams improve brand performance in AI search. More than 3,000 agencies and brands, including global enterprises, rely on Peec AI to understand where they stand, what shapes how they appear and get recommended, and where to act. The company combines original research with rigorous measurement of the markets, models, sources, competitors, and websites behind AI answers, helping teams separate meaningful signals from noise and make better-informed decisions about what to do next. Peec AI is one of Europe's fastest-growing AI software companies, surpassing $15M in ARR since launching in January 2025. Founded by Marius Meiners, Daniel Drabo, and Tobias Siwonia, Peec AI has raised $40M from investors including Singular, Antler, Combination VC, identity.vc, and S20. Learn more at peec.ai. About Ramp Rate Ramp Rate is a vendor directory powered by aggregated, anonymized data from Ramp's corporate card and bill pay platform, benchmarking corporate spend across billions of dollars and tens of thousands of businesses. Press Inquiries Legal Disclaimer: EIN Presswire provides this news content "as is" without warranty of any kind. Africa Finance Today do not accept any responsibility or liability for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this article. If you have any complaints or copyright issues related to this article, kindly contact the author above.
Meet Cloudflare Clef-omni: A 30B-A3B multimodal decision model for audio, video, image and text. October 11, 2026 Cloudflare has released Clef-omni, a new open-weight member of its Clef decision model family. Clef-omni is a Cloudflare decision model that scores text, images, audio and video in 1 call. It arrived on October 9, 2026, 8 days after the original Clef and Clef-flash launch. The same update cut Clef-flash pricing below TypeSafe's Jev and made hosted Clef roughly 2x faster. For developers, this removes the transcription and captioning steps that multimodal agent pipelines usually need. Tl;dr. * Size: 30B total, 3B active parameters (MoE); 64,000-token context on Workers AI * Runs on: Workers AI as @cf/cloudflare/clef-omni; self-hosting tested on a single H200, about 64 GB GPU memory in bfloat16 * Performance: Close to Clef on routing and intent tasks, behind it on most reasoning and workflow evals * Best: 97.7 macro-F1 on CLINC150+OOS, top of the 4 Clef and Jev models * Worst: 2.6 accuracy on POP909-CL, and 47.4 on GPQA Diamond versus Jev's 78.3 * Bottom line: * Best: one call decides across 4 modalities. * Worst: weaker accuracy than text-focused Clef on hard reasoning. What is a decision model? A decision model returns typed answers with probabilities instead of generated text. You send a state and a schema of questions. The model returns a probability for every allowed option. Clef supports 3 question types: noul (true or false), choice and score. The format follows the Jev and SystemOne API from TypeSafe AI. How does Clef-omni work? Clef-omni is post-trained from Qwen3-Omni-30B-A3B-Instruct. Cloudflare keeps the comprehension backbone and drops the text-to-speech output path. The model runs a single prefill pass over the full payload. Video frames and audio are synced into 1 sequence for joint processing. A small joint schema head then scores the options. Each valid option first gathers evidence from the input. Field vectors then cross-attend across the full context to compute confidence scores. Cloudflare froze the Qwen3 backbone and trained LoRA adapters. Training combined label-smoothed cross-entropy with Brier score calibration. Supported inputs on Workers AI, per the developer docs: * Images: up to 4, PNG, JPEG or WebP, 64 to 1,024 tokens each * Audio: up to 4 clips, WAV or MP3, 300 seconds each, about 780 tokens per minute * Video: up to 2 clips, MP4 or WebM, 60 seconds each, sampled at 2 frames per second * Questions: 1 to 64 per request Cloudflare reports text decisions in about 130 ms at the median and images in about 150 ms. A 21-second video with sound scores in about 1.5 seconds. How does Clef-omni perform on benchmarks? Cloudflare compared Clef-omni against Clef, Clef-flash and Jev on its Decision Index run. Clef-omni leads on intent classification: 94.8 macro-F1 on BANKING77 and 97.7 on CLINC150+OOS. It also tops Amazon ESCI at 57.8. It trails elsewhere. Home appliances case-exact accuracy is 69.3, versus 97.73 for Clef-flash. When2Call accuracy is 63.3, versus 80.97 for Jev. On TypeSafe's workflow evals, Clef-omni scores below Clef in all 5 reported metrics. Clef-omni vs Clef vs Clef-flash vs Jev. Benchmark scores are from Cloudflare's announcement and were not independently verified. How much faster is Clef? Cloudflare moved hosted Clef to SGLang and tuned its serving stack. No new weights were released. Median latency fell from 262 ms to 152 ms at about 800 tokens, a 1.7x gain. At about 3,400 tokens it fell from 616 ms to 305 ms, a 2.0x gain. At about 16,000 tokens it dropped from 2,721 ms to 1,635 ms. Clef support lands in SGLang 0.5.22. Key takeaways. * Clef-omni scores text, images, audio and video in 1 decision call. * It is a 30B-A3B MoE on Qwen3-Omni, Apache-2.0, at $0.15 per million tokens. * Clef-flash now costs $0.038, under Jev's $0.042, with 24k hosted context. * Hosted Clef is 1.7x to 2.0x faster at the median after moving to SGLang. * Clef-omni wins intent tasks but trails Clef on reasoning and workflow evals.