Summer 2026
AI Inference Intern
Posted on 2/12/2026
Perplexity AI
1,001-5,000 employees
Advanced answer engine delivering source-backed replies
No salary listed
London, UK
Hybrid
Three days from the London office and two days working from home each week. UK work visa sponsorship is unavailable.
Master's, PhD
 Software Engineering (1) |
|---|
Get referred to Perplexity AI
See people who can refer or advise you
- Candidates must be pursuing a Master's or PhD in Computer Science with a focus on performance-related subjects such as High-Performance Computing, compilers, or distributed systems.
- Candidates must have a strong engineering track record with proven knowledge of fundamentals and programming languages, including multi-threaded programming, networking, compilation, and systems programming.
- Candidates must have experience with machine learning frameworks such as Torch and JAX.
- Candidates must have experience with GPU programming using CUDA and Triton.
- Candidates must have experience with High-Performance Computing using OpenMPI.
- Work with the inference team to improve serving latency and throughput.
- Bring up support for new models and state-of-the-art inference optimizations or quantization schemes.
- Optimize inference across the entire stack, from GPU kernels to serving endpoints.
Perplexity AI provides an answer engine that delivers precise, reliable responses to user questions by using up-to-date sources. It serves individuals who want quick answers and businesses that need detailed information, drawing on current data and source links to back its results. The product works by retrieving information from reputable sources, compiling concise answers, and presenting citations to ensure trustworthiness. It differentiates itself from competitors by prioritizing current, source-backed information and offering access that spans personal and enterprise use, potentially supported by subscriptions, advertising, and partnerships. The company's goal is to be a dependable tool for immediate, accurate information across a wide range of queries.
Company Size
1,001-5,000
Company Stage
Private
Total Funding
$1.8B
Headquarters
San Francisco, California
Founded
2022
Get referred to Perplexity AI
See people who can refer or advise you
Simplify's Take
What believers are saying
- July 2025 Airtel bundle drove 56 million India downloads and July 2026 revenue growth.
- March 11, 2026 Ask conference expanded Computer, enterprise connectors, and multi-model orchestration for customers.
- Snap's 2026 partnership and Canary Data integration deepen distribution and enterprise research workflows.
What critics are saying
- August 3, 2026 Reuters: Reddit's copyright suit survived dismissal; discovery threatens training and attribution practices.
- March 2026 Amazon litigation over Comet agent access targets Perplexity's core agentic-browser product.
- CNN sued Perplexity in 2026 over content distribution; adverse rulings or licensing costs crush margins.
What makes Perplexity AI unique
- Perplexity indexes web sources, then synthesizes answers across multiple frontier models in one interface.
- Comet and Computer turn search into action, not just chat, across browsers and workflows.
- July 29, 2026 Numbat and May 13, 2026 security work show serious agent-security depth.
Help us improve and share your feedback! Did you find this helpful?
Benefits
Health Insurance
Dental Insurance
Vision Insurance
401(k) Retirement Plan
Company Equity
Company News
Perplexity's year-long free AI experiment in India offers early clues about converting giveaway users into paying customers. The company partnered with telecom giant Airtel in July 2025, offering 360 million customers free 12-month Pro subscriptions worth $200. The campaign drove 56 million downloads over seven months, more than nine times the prior period, according to Sensor Tower data shared with TechCrunch. Monthly active users peaked at 22 million in October. As free subscriptions expired, downloads fell 90% between February and July. However, monthly active users remained at 14 million in July, still five times pre-promotion levels. In-app revenue rose 60% despite declining downloads. From mid-July through mid-August, average daily revenue increased 27% above the first half of 2026. Whether this reflects deliberate conversions or failed cancellations remains unclear, as subscriptions auto-renewed unless users opted out. The experiment preceded similar moves by OpenAI and Google in India, a market crucial for AI adoption but difficult to monetise.
Canary Data and Perplexity have partnered to integrate Canary's investment research platform into Perplexity's answer engine and Perplexity Computer. The partnership allows joint clients to access Canary's proprietary datasets, contextual layer, and AI analysts directly within Perplexity's environment. Canary combines dozens of proprietary datasets — including undisclosed business events, AI disruption exposure, and investor idea pitches — with alternative data such as credit card transactions and web traffic. AI adds context to make the data useful for investors. Perplexity users can now query Canary's datasets, work with contextualised versions, or request finished analysis from Canary's AI analysts. The companies are launching with a bring-your-own-licence model, enabling clients to connect existing Canary entitlements to Perplexity without additional API configuration. Canary was founded by a former Tiger Global partner and is backed by Tiger Global and Arena Holdings.
Firebird launches Nvidia-powered 300MW AI data centre in Armenia. - John Tanner - 11 August 2026 - 2940 views Nvidia announced on Saturday that AI cloud startup Firebird has launched a new AI data centre in Armenia, powered by Nvidia GPUs and Dell Technologies' AI infrastructure. Firebird's AI data centre in Hrazdan - which was delivered in six months - was built on Nvidia's AI factory platform, which includes reference architecture, GPUs, networking and AI software. The facility hosts over 70,000 Nvidia Rubin and Blackwell GPUs on Dell PowerEdge servers, with its AI infrastructure supporting an IT load capacity of 300MW. The data centre also runs on Nvidia's DSX platform, which integrates computing, networking, power and cooling as one codesigned system. Firebird said that enables it to run up to 40% more GPUs on the same footprint, producing more tokens per dollar and extracting more value from every megawatt of capacity Schneider Electric provided the power infrastructure supporting the AI data centre, including medium- and low-voltage switchgear, three-phase uninterruptible power supply systems and rack enclosures to make sure the power buildout keeps up with pace of compute. Meanwhile, Vertiv has provided a cooling architecture combining chilled-water technology, advanced controls and Vertiv TrimCooler technology for efficient heat rejection. The cooling resources are centrally coordinated by Vertiv's chilled water manager to improve visibility, efficiency and responsiveness as demand shifts with AI workloads. Nvidia said Perplexity AI has already signed on as a customer, using Firebird's high-performance AI infrastructure for its AI agent platform and answer engine. Nvidia said the Hrazdan AI data centre is the largest in the CIS, and will help make the region a magnet for global companies building and running AI, as well as a catalyst for local developers, researchers and enterprises. Firebird co-founder Alexander Yesayan said the Armenian data centre is part of its broader ambition to pursue an AI infrastructure roadmap spanning Armenia, Kazakhstan and additional markets. "Our ambition for what we are building in the next two years or so is roughly 2 GW of capacity around the world," Yesayan said. "We're very focused on merging into frontier markets." Firebird also announced that Nvidia intends to invest in the company, following an earlier investment by CoreWeave this year, both of which will help finance that expansion plan. No details were given on the size of the Nvidia investment.
Firebird opens major AI factory in Armenia. Firebird has launched what it calls the CIS region's largest AI factory in Armenia, creating a computing hub powered by NVIDIA accelerated computing and Dell Technologies AI infrastructure. The opening was attended by Armenian Prime Minister Nikol Pashinyan, Kazakhstan Deputy Prime Minister Zhaslan Madiyev and U.S. chargé d'affaires David Allen. The facility is designed to give developers, startups, enterprises, universities and public institutions local compute capacity to train, fine-tune and deploy AI models for Armenian languages, industries and national priorities. Firebird plans to deploy more than 70,000 NVIDIA Rubin and Blackwell GPUs and 300 megawatts of AI infrastructure capacity in Armenia by the end of 2027. The build uses the NVIDIA DSX platform to integrate accelerated computing, networking, power and cooling, with Firebird saying the design can run up to 40% more GPUs on the same footprint. Schneider Electric is supplying power infrastructure, while Vertiv is providing chilled-water cooling systems and controls. Firebird is also pursuing an approximately 2-gigawatt AI infrastructure roadmap across Armenia, Kazakhstan and additional markets, with NVIDIA intending to invest after an earlier CoreWeave investment. Early demand includes Perplexity, which is working with Firebird on high-performance infrastructure for its AI agent platform and answer engine. Originally reported by blogs.nvidia.com Read the source
Amazon loses US court ban on Perplexity's AI shopping tools. 05 Aug 2026 03:53AM (Updated: 05 Aug 2026 04:23AM) Add CNA as a trusted source to help Google better understand and surface our content in search results. Aug 4: A U.S. appeals court on Tuesday overturned a ruling that had temporarily barred Perplexity from using its AI-powered agentic shopping tools on Amazon's platform. In a novel legal dispute, the San Francisco-based 9th U.S. Circuit Court of Appeals determined that Amazon was unlikely to succeed on its claim that Perplexity's AI agents violated a federal computer-hacking law. The ruling is the first by a federal appeals court to address whether AI agents acting on behalf of users can legally access online platforms, a question with broad implications for the fast-growing field of agentic AI tools that can browse, shop and transact across the web. "We respectfully disagree with today's decision on the preliminary injunction," an Amazon spokesperson said. "We remain confident in our case and are evaluating our next steps." "Perplexity will continue to fight for the right of internet users to choose whatever AI they want," company spokesperson Jesse Dwyer said in a statement. "We have always been confident the truth will prevail and the rights of users will not be chipped away." Agentic AI systems can plan, reason and execute tasks with limited human oversight. Amazon sued Perplexity in November, accusing the AI startup of covertly accessing private Amazon customer accounts through its Comet browser and associated AI agent, which can log into users' online shopping accounts and place orders on their behalf. The lawsuit said Perplexity's system posed security risks and that it had ignored repeated requests to stop. Perplexity responded that the lawsuit lacked merit and was a "bald attempt" to block Amazon users from using Comet because AI agents "don't have eyeballs to see the pervasive advertising Amazon bombards its users with." A federal judge in California temporarily blocked Perplexity in March from using its agentic AI on Amazon's platform, finding Amazon had presented "strong evidence" that Perplexity's AI agent broke a U.S. computer fraud and abuse law. The 9th Circuit overturned the ban on Tuesday, favoring Perplexity's argument that it did not violate a statute that prohibits people from accessing computers and obtaining information from them without permission. The appeals court said it was Perplexity's users, using the company's AI agents, who "accessed" Amazon's platform under the law, rather than Perplexity itself.