Software Engineer
About the job
- At least 4 years of professional software engineering experience; strong junior and mid-level candidates with a track record of shipping are also considered.
- Backend or full-stack engineering skills, including experience designing and building scalable and reliable distributed systems that serve high traffic and a large user base.
- Ability to take ambiguous product needs, break them down, and ship durable systems with clear customer impact.
- Experience building product or platform features requiring careful data modeling, state management, backend correctness, and high-quality user experience.
- Ability to make complex administrator, billing, and permissions workflows simple and intuitive.
- Self-motivation, strong ownership instincts, and the ability to identify problems, propose solutions, and drive improvements independently.
- Genuine interest in artificial intelligence products and willingness to learn quickly.
- Build intelligent products across onboarding, membership management, organization insights, permissions, subscriptions, billing, and usage controls.
- Build systems that help organizations manage Perplexity at scale, from initial setup through expansion, enabling administrators to onboard teams, configure features, manage usage, and understand spending.
- Collaborate with Product Management, Design, Data Science, Sales, and enterprise customers to turn complex administrative, billing, and governance needs into simple, reliable product experiences.
- Own critical platform primitives end-to-end, including frontend administrator flows, backend services, data models, billing integrations, observability, and production operations.
- Experience building enterprise Software as a Service products, administrator consoles, membership systems, role-based access control or permissions, or organization management features.
- Experience with subscriptions, billing systems, or usage-based pricing.
- Familiarity with security, compliance, or governance concerns in enterprise products.
- Experience partnering directly with Sales, Finance, Support, or enterprise customers to solve high-impact product and platform problems.
- Time spent at a fast-growing startup or on a high-ownership engineering team.
About the company
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
- Microsoft's January 29, 2026, $750 million Azure deal anchors infrastructure and model supply.
- Fast Search delivered 64.3% benchmark score and 68% lower estimated cost on September 30.
- AMD's September 2026 Portable Computer integration expands Perplexity into local AI workflows.
What critics are saying
- DaVoice sued on September 25, 2026; Reddit, CNN, New York Times, and Amazon threaten revenue.
- Amazon won March 10, 2026, blocking Perplexity shopping agents; appeal risk remains acute.
- If courts restrict scraping or agents, Perplexity loses the data moat powering answers and Comet.
What makes Perplexity AI unique
- Photon cut production p99 latency from 800 ms to 65 ms on September 24, 2026.
- Computer now bundles Skills Marketplace, Effort Mode, and local execution across Macs and RTX PCs.
- Microsoft Foundry gives Perplexity frontier-model access from OpenAI, Anthropic, and xAI.
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 details Photon, its search retrieval engine. Perplexity says Photon handles retrieval and ranking in its production search stack. Its reported 65 ms p99 covers Photon's stages; Fast Search is a separate API mode. September 30, 2026 Key points * 01 Perplexity describes Photon as the internal engine that retrieves and ranks candidate web pages for its production search. * 02 Perplexity reports a production p99 of about 65 ms for Photon's stages, compared with about 800 ms for the previous retrieval-and-ranking system. * 03 Fast Search is a separate Search API mode built on Photon, with reported single-call latency of 160 ms p50 and 230 ms p95. * 04 Across six benchmarks and 3,554 selected tasks, Perplexity reports a 64.3% score for Fast Search and 64.0% for its default preset. * 05 Fast Search costs $1 per 1,000 successful API requests, according to Perplexity's documentation. Perplexity says Photon handles search retrieval and ranking at a 65 ms p99. See what that figure covers and how Fast Search's API metrics differ. On September 24, 2026, Perplexity announced the Fast Search option for its Search API and described Photon, the internal retrieval-and-ranking engine it says powers its production search. Perplexity reports a p99 latency of about 65 milliseconds for Photon's stages, down from about 800 milliseconds in its previous system. Fast Search is a separate API mode built on Photon, with its own latency and benchmark figures. What Photon does in Perplexity search. Photon retrieves candidate web pages and ranks them for Perplexity's search stack; it is not a standalone consumer search engine. Perplexity says the system replaced an adapted open-source engine as its index and workloads grew, and places Photon within its broader move to in-house Rust-based search infrastructure. How a query moves through Photon. A request passes from a load balancer to Photon's broker, which selects a group of shards. The shards retrieve and rank candidate pages; the broker then merges those candidates and fetches key fields for selected documents before returning results to the higher-level search system. Photon uses an inverted index, which maps terms to the documents that contain them. It stores compact per-document ranking records called "docblobs" and uses different representations for short and long posting lists. A WAND-like traversal uses score bounds to limit which candidates need further ranking. How Perplexity builds and serves its index. Perplexity separates index construction from live query serving. Pillar prepares source data in YTsaurus tables, and indexers build shard structures from that data. Versioned index files are stored for deployment; a controller rolls new index versions out across serving groups. That separation also shapes the reported operating figures: Perplexity says Photon uses about 20% fewer equivalent serving machines than the previous system and stores about 2.5 times as much data per document. The company says it can build its full web index in a single-digit number of hours. What Photon's reported p99 latency measures. Perplexity's production p99 comparison covers the retrieval-and-ranking system: about 800 ms for the previous engine and about 65 ms for Photon. The 65 ms figure covers Photon's internal stages and excludes later stages in the higher-level search stack. | System | Reported production p99 | Measurement scope | | Previous retrieval-and-ranking system | About 800 ms | Retrieval and ranking in Perplexity's production context | | Photon | About 65 ms | Photon's internal stages; later search-stack stages are excluded | Fast Search benchmarks and API access. Fast Search combines Photon with a ranking configuration tuned for agent workflows. Perplexity reports 160 ms p50 and 230 ms p95 for a single Fast Search API call. Those figures describe API-call latency, not Photon's production p99. The p50 is the median; p95 is the 95th-percentile latency. Across six benchmarks and 3,554 selected tasks, Perplexity reports a 64.3% task score for Fast Search and 64.0% for its default preset. The company's estimated model-plus-search costs for that benchmark set were $59.73 and $187.60, respectively. These are estimated costs across the tasks, not per-request API prices. | Perplexity-reported measure | Fast Search | Default preset | Context | | Task score | 64.3% | 64.0% | Six benchmarks; 3,554 selected tasks | | Estimated model-plus-search cost | $59.73 | $187.60 | Same benchmark task set | | Long-tail DCG relevance | 2.21 | 2.45 | Separate internal tests | | Answer availability | 0.567 | 0.596 | Separate internal tests | The long-tail results put the small benchmark-score difference in context: Perplexity reports lower DCG relevance and answer availability for Fast Search than for its default preset in separate internal tests. To call Fast Search, the Search API uses search_type: "fast" on POST /search; the max_results setting can range from 1 to 20. Perplexity's documented tariff is $1 per 1,000 successful requests. Other Languages Read this article in another available language.
Perplexity introduces Photon: A rust-based retrieval engine that cuts p99 latency from 800 ms to 65 ms. September 30, 2026 Perplexity has released Photon, an in-house retrieval and ranking engine written in Rust. It replaces an open-source engine Perplexity had forked for its AI-native search stack. Photon now handles retrieval and ranking for all production traffic. It also powers a new Fast Search mode in the Perplexity Search API. Perplexity reports single-call latency of 160 ms at p50 and 230 ms at p95. Is it deployable? Yes, as a hosted API. Set search_type: "fast" on POST /search and pay $1 per 1,000 requests. Photon itself is not open source, so the engine cannot be self-hosted. Why Perplexity replaced its old engine. The old engine hit 3 limits as the index grew: * Tail latency: Production p99 sat near 800 ms. The dataset exceeded RAM, so mlock was not an option. Cold reads triggered major page faults that stalled queries. * Merge spikes: During disk index fusion, p99 climbed to about 1.2 s for 10 to 15 minutes. * Slow recovery: Deploying and syncing an extra cluster could take more than a week. Recovery also raised the share of partial responses. Perplexity team concluded that building from scratch was simpler and cheaper than maintaining its fork. How Photon works. A load balancer routes each request to a Photon broker. The broker fans out to a shard group and watches for timeouts. Each shard runs retrieval, initial ranking, and second-stage ranking. The broker then merges candidates and fetches key document fields. * Adaptive posting lists: Short lists sit inline within a single page. Longer lists split into blocks of fixed document ID ranges. Sparse blocks store sorted offset arrays and use galloping search. Dense blocks use bitmaps, so membership becomes a single bit lookup. * Budgeted traversal: A WAND-like algorithm splits lists into driving lists and probe lists. Cheap presence checks bound each candidate's maximum score first. Exact term frequencies are read only when a candidate can clear the threshold. * Docblob records: Each document gets a compact record of frequencies, field masks, and positions. Terms use Elias-Fano encoding, so ranking decodes only the matched terms. Ranking a candidate needs just 1 lookup per document. * Batched async reads: Record offsets are known upfront, so disk reads go out in batches through io_uring. The cache checks the whole batch first. Readers take no locks, and eviction uses CLOCK instead of a shared LRU list. * Separate build and serve: Indexers build versioned shard indexes from YTsaurus tables on dedicated nodes. A controller rotates serving groups one at a time and warms caches with replayed search-log queries. A full web index now builds in a single-digit number of hours. Interactive explainer: inside Photon. Production results. * p99 retrieval and ranking latency fell from about 800 ms to about 65 ms. This covers Photon's stages only. * Photon runs on about 20% fewer serving machines than the old content nodes. * It stores about 2.5x as much data per document, which Perplexity used to improve ranking quality. * Pinning the same dataset with mlock would need an estimated 4.6x the resident memory Photon uses today. * Index version switches no longer cause latency spikes. Fast Search: speed and cost for agents. Fast Search pairs Photon with lighter ranking tuned for agentic workflows. Perplexity tested it on 6 benchmarks: WideSearch, BrowseComp, DSQA, FRAMES, SEAL-0, and SEAL-Hard. Across 3,554 tasks, Fast scored 64.3% at $59.73 in estimated model-plus-search cost. The default preset scored 64.0% at $187.60, so Fast was about 68% cheaper. The trade-off shows up in broader search quality. On internal long-tail benchmarks, relevance (DCG) fell from 2.45 to 2.21. Answer availability dropped from 0.596 to 0.567, a loss of 2.9 percentage points. Perplexity recommends Fast for day-to-day agent loops and the default for hard, ambiguous queries. curl -X POST 'https://api.perplexity.ai/search' \ -H "Authorization: Bearer $PERPLEXITY_API_KEY" \ -H 'Content-Type: application/json' \ -d '{"query": "latest stable Rust release", "search_type": "fast", "max_results": 5}' On Python SDK 0.43.4 and 0.43.5, pass extra_body={"search_type": "fast"} per the docs. Fast Search vs closest competitors. All latency figures are vendor-reported under different setups, so they are not like-for-like. Key takeaways. * Photon is Perplexity's Rust retrieval and ranking engine, now serving all production traffic. * Production p99 latency dropped from about 800 ms to about 65 ms. * Fast Search reports 160 ms p50 and 230 ms p95 at $1 per 1,000 requests. * Fast cut estimated agent task cost by about 68% at comparable task quality. * It trades some retrieval relevance, so keep the default preset for hard queries. Need to partner with Marktechpost LLC. for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.? Connect with Marktechpost LLC. Asif Razzaq is the CEO of Marktechpost AI Media Inc... As a visionary entrepreneur and engineer, Asif is committed to harnessing the potential of Artificial Intelligence for social good. His most recent endeavor is the launch of an Artificial Intelligence Media Platform, Marktechpost, which stands out for its in-depth coverage of machine learning and deep learning news that is both technically sound and easily understandable by a wide audience. The platform boasts of over 2 million monthly views, illustrating its popularity among audiences.
Perplexity AI sued by startup for allegedly stealing 'wake word' technology. The complaint says that Perplexity stole DaVoice's secrets after the companies collaborated on unspecified technology "built with DaVoice's proprietary source code" By reuters. Updated - September 26, 2026 at 12:45 PM. FILE PHOTO: Perplexity AI application icon is seen in this illustration taken January 4, 2024. REUTERS/Dado Ruvic/Illustration/File Photo | Photo Credit: Dado Ruvic Startup DaVoice has sued Perplexity AI in California federal court for allegedly stealing trade secrets related to "wake word" technology for AI assistants. * The heavily redacted complaint, filed on Thursday, said that Perplexity stole DaVoice's secrets after the companies collaborated on unspecified technology "built with DaVoice's proprietary source code, inference logic, network architecture, training approaches, training data, and know-how." * Spokespeople for the companies did not immediately respond to requests for comment on Friday. * DaVoice specializes in technology that allows devices to continuously listen for "wake words" to activate them. * The startup said that it has collaborated with companies including Volkswagen and internet-of-things company Samsara. * Perplexity is best known for its AI-powered search engine technology and Comet browser and AI assistant The case is DaVoice v. Perplexity AI Inc, US District Court for the Northern District of California, No. 3:26-cv-10909. Published on September 26, 2026 THIS AD SUPPORTS OUR JOURNALISM. SUBSCRIBE FOR MINIMAL ADS.
Startup DaVoice sues Perplexity AI over stolen wake word technology. Voice-AI startup DaVoice has sued Perplexity AI in federal court, alleging the company stole its wake word technology, including source code and training data, after the two companies collaborated. Perplexity calls the claim baseless and says their agreement protected its right to build competing tech. A Swiss voice-AI startup says Perplexity took its wake word code, its training data and its architecture, then kept building without it. Perplexity says the contract let it do exactly that. DaVoice, the voice-AI startup that does business under the legal name SyteMLLabs, filed a heavily redacted complaint on September 24 in the U.S. District Court for the Northern District of California, accusing Perplexity AI of stealing trade secrets tied to wake word technology, the software that lets a device sit there listening for a phrase like "Hey Siri" without streaming every sound it hears to the cloud. The case is DaVoice v. Perplexity AI Inc., No. 3:26-cv-10909, and DaVoice is represented by Davida Brook and Ian Crosby of Susman Godfrey, according to Courthouse News Service. The two companies had worked together. That's the part that makes this messier than a typical copycat claim. DaVoice, co-founded by Ofer Fryman and Helge Voss, says it entered into confidential agreements with Perplexity to license its wake word system, technology the startup says runs at above 90% accuracy and undercuts rivals on price. According to the complaint, Perplexity absorbed DaVoice's proprietary source code, inference logic, network architecture, training approach and training data during that collaboration, then allegedly built its own version using what DaVoice calls the false and misleading promise of a lasting, lucrative partnership. Perplexity isn't backing down. Jesse Dwyer, the company's chief communications officer, called the suit "a baseless attempt to claim ownership of standard wake-word technology," and said the agreement DaVoice signed expressly protects Perplexity's right to develop its own "similar, equal or competitive products or services." In other words: Perplexity's defense isn't that it didn't build a competing product. It's that the contract said it could. Wake word detection sounds like a small piece of engineering, but it's the gatekeeper for every voice assistant that exists. Get it wrong and the device either ignores you or listens to everything, which is its own privacy problem. Get it right and you've built the front door to whatever the assistant does next. Perplexity has spent 2026 pushing hard into that territory, expanding its Comet browser's voice features and its assistant mode, and a startup with real automotive and IoT customers, DaVoice has worked with Volkswagen and IoT company Samsara, is exactly the kind of technology a fast-moving AI company might want to shortcut rather than build from scratch. Anthropic's full threat report reveals Xiaomi wasn't acting alone: six other Chinese AI labs, including Alibaba, DeepSeek and Moonshot, ran similar campaigns to extract Claude's outputs, with Moonshot caught secretly routing live user queries through Claude without disclosure. - chinese AI labs copying Claude without attribution - how Anthropic caught multiple companies stealing Claude That's the pattern showing up across the industry this year. Smaller, specialized firms build a hard piece of infrastructure, hand it to a bigger platform under an NDA or a pilot deal, and then find themselves competing against something that looks a lot like their own work. Whether that's illegal depends entirely on what the paperwork actually says, and paperwork is exactly what this case will turn on. DaVoice's complaint is redacted precisely because the underlying agreements, and what they do or don't permit, are the whole fight. Frankly, both sides can't be right about what that contract says, and a judge in San Francisco is going to have to read it line by line to find out. If Perplexity's version holds, this becomes a cautionary tale for startups who license their core technology to a much larger partner without locking down what happens after the relationship sours. If DaVoice's version holds, it becomes a cautionary tale for platforms that treat a vendor's source code as raw material once the ink is dry. Perplexity hasn't filed a formal answer to the complaint yet. The case now sits in the Northern District of California, where trade secret disputes involving AI companies have become a fairly regular occurrence this year, and where the next filing will likely be Perplexity's motion to dismiss or its own answer laying out exactly what that licensing agreement permitted. Join the discussion. No replies yet. Start the discussion. Dave Barr is a professional Marketing Strategist With Over 6 Years Of Experience in PR. His primary area of expertise is public relations and social branding. Dave has been associated with various content projects from across the world on a regular basis. He has also had associations with big and reputed news networks. Dave contributes to Startup Fortune in the Business, Marketing and Technology sections.
Advanced micro devices (AMD) stock gains on Perplexity AI collaboration. Key highlights. Table of Contents * Shares of AMD advanced 2% Friday following the reveal of innovative AI capabilities developed with Perplexity. * The collaboration integrates Perplexity's Portable Computer into Windows machines equipped with AMD Ryzen AI Max processors. * PC owners can execute AI operations locally on their devices, preserving Perplexity Computer credits for other uses. * This development extends AMD's agentic PC initiative that gained momentum in 2025. * Analysts maintain a Strong Buy consensus on AMD with a mean price target of $648.07. Shares of AMD experienced a 2% uptick Friday after the semiconductor manufacturer revealed a collaborative AI tool development with Perplexity. The positive market response reflected investor enthusiasm for AMD's continued expansion in AI-enhanced computing devices. At the heart of this announcement is Perplexity's Portable Computer technology. The tool is now compatible with Windows computers powered by AMD's Ryzen AI Max processors, which includes the Ryzen AI Halo developer platform. This integration enables PC users to execute artificial intelligence operations directly on their hardware. Users can establish automated recurring tasks without consuming Perplexity Computer credits for processing. This latest development represents the culmination of AMD's multi-year vision. The company introduced its initial AI-capable PCs in 2023, featuring an integrated neural processing unit within an x86 architecture chip. The following year, 2025, saw AMD accelerate its commitment to agentic computing. These advanced systems enable AI agents to operate locally on the device instead of depending exclusively on cloud infrastructure. AMD leadership weighs in on innovation. AMD executive Jack Huynh addressed the partnership announcement mid-week. Huynh serves as senior vice president and general manager of AMD's Computing and Graphics Group. "We're entering a new era of computing, where personal AI transforms the PC from a tool into an intelligent partner," Huynh said. He emphasized that computers are evolving into systems capable of understanding, reasoning, and acting in collaboration with users. According to Huynh, the Ryzen AI Max processor family enables AI models and intelligent agents to execute directly on personal computers. This architecture provides users the autonomy to work locally while maintaining access to more powerful cloud-based processing when circumstances require it. The processor design merges CPU and GPU capabilities with an expansive unified memory architecture. This configuration enables local AI model execution to run concurrently with standard applications without performance degradation. Perplexity contributes its model infrastructure and agent technology to the partnership. The companies jointly assert that this collaboration transforms agentic PCs from a specialized concept into a viable solution for mainstream consumers. Analyst sentiment on AMD stock. The strategic timing of this product launch carries significance. While cloud-based infrastructure currently dominates AI processing, AMD is positioning itself for a shift toward local task execution as user preferences evolve. This forward-looking strategy explains Friday's market enthusiasm. Should local AI adoption accelerate as AMD anticipates, the company stands to gain substantial advantages from its early-mover position in this segment. Financial analysts maintained an optimistic outlook on AMD stock even before this announcement. The consensus rating stands at Strong Buy, supported by 30 Buy recommendations and six Hold ratings compiled over the previous three-month period. The average analyst price target for AMD stock reaches $648.07. Notably, zero Sell ratings exist among the tracked analyst community. The Perplexity collaboration provides additional validation for investors monitoring AMD's artificial intelligence PC roadmap. Friday's stock appreciation demonstrates positive market reception to this strategic development. Limited Time Offer Get 3 free stock ebooks. Discover top-performing stocks in AI, Crypto, and Technology with expert analysis. * Top 10 AI Stocks - Leading AI companies * Top 10 Crypto Stocks - Blockchain leaders * Top 10 Tech Stocks - Tech giants