November 9, 2022

Meet Vectara: powerful, free neural search

Posted in Category: Work — Amr Awadallah @ 12:39 am | link | | Comments Off on Meet Vectara: powerful, free neural search

We’re in a transformative moment in computer science. AI breakthroughs like GPT-3, Copilot, DALL-E and Stable Diffusion are changing how computers understand, interpret, and interact with human language. These Large Language Models have already shown they can author text, write code, create art, and remodel knowledge in ingenious new ways.

Yet these breakthroughs aren’t evenly distributed. Sophisticated AI technologies like neural networks are underpinning services we use every day, but they’re not accessible to the majority of developers who could use them to enhance user experience in their own applications. 

This is a huge problem to solve, and I’m privileged to lead a brilliant team at Vectara that is taking on this challenge with our first product, which we’re making freely available for production workloads today. As the incredible potential and wide range of applications for neural networks are coming into view, we’re starting our journey with an offering that overcomes the limitations and difficulties of keyword search.

We’ve built a novel neural search product that is simple for companies to integrate into their websites and applications. The Vectara platform empowers organizations to easily and quickly embed the most advanced natural language processing available into their apps and site search, delivering incredibly relevant results with seamless ease of use for developers.  

Vectara’s neural search is the most effective way to achieve semantic search. The IT industry continues to emulate semantic understanding with keyword search algorithms that haven’t been updated in decades. The results continue to be underwhelming. In contrast, neural search is an unqualified, sea-change breakthrough for search. As an experienced purveyor of breakthrough technologies (see Cloudera and “Big Data”), I am convinced that nearly every search application will be powered by native neural search within the next 5 years. 

Breakthrough neural search as a service

Search is essential for navigating the world’s knowledge, whether combing through technical materials or seeking guidance on life’s big questions. However people ask – with shorthand, typos, or using voice – they want fast, accurate results. Yet the algorithms that power most search applications rely on keywords, leading to narrow results which often miss the intent behind the question. 

Very few search experiences today can answer queries based on contextual semantic understanding. The technology that makes this feat possible, which uses large language models and vector similarity as a first principle in information retrieval, has been expensive and difficult to deploy without extensive machine learning expertise. With our free,   easy-to-use neural search as a service platform, Vectara overcomes both of these barriers and opens the door to unprecedented improvements in search relevance

With this initial offering, we aim to solve the issues that have plagued keyword search engines for over 30 years. These algorithms require intense constant gardening to apply sophisticated language rules to emulate even the most basic levels of semantic relevance. We also recognize that users increasingly demand answers delivered in their language of choice and with the appropriate cultural or business context. By leveraging the power of AI to identify implied meanings, Vectara makes it possible for every user to be heard and understood.  

Unparalleled search performance for all

Our platform applies deep vector encoding for all indexed content and all issued queries, enabling a true meaning-to-meaning response that retrieves relevant information based on neural-powered understanding of intent and context. Yet even the most advanced models are only useful if they can be easily applied to real-world challenges and integrated into existing workflows. That’s why we built our developer-first platform with powerful yet easy to use APIs, giving any organization access to cutting-edge AI without the high cost or uncertainty associated with other solutions. 

Vectara delivers a complete search pipeline from extraction and indexing to retrieval, rerank, and calibration. As a composable neural data processing platform, developers can easily integrate Vectara into their stack to quickly embed or enhance site and app search, product discovery, customer support, user navigation, and many other possibilities. With our platform, searches are more accurate and less frustrating, chatbots get smarter, product recommendations become more personalized, and users experience more delight engaging with content and applications. 

Simple to integrate

Vectara’s neural search as a service is one of the most powerful applications of neural large language models available today. Unlike other search as a service offerings, however, no initial or continual tuning or training is required. Vectara uses large language models to analyze data without additional training or the continuous investment in updating language configurations required by keyword platforms such as stop words, synonyms, acronyms, emojis, spelling corrections, part-of-whole relationships, lemmatization and stemming rules. 

We engineered Vectara so that even the smallest teams can implement a highly available, auto-scaling, and secure, enterprise-grade service. We carefully monitor system latency, and our API-first platform to allow you to query and index from anywhere. No specialized search engineering or AI/ML knowledge is necessary to integrate this advanced neural search into your site or applications. Of course Vectara’s APIs allow opinionated search developers to customize the behavior at a deeper level if desired.

A complete ML search pipeline

Vectara provides a low latency, highly fault-tolerant platform that delivers a complete, yet composable, ML search pipeline from extraction, encoding, and indexing to retrieval, and rerank. Vectara automatically routes data across relevant backend microservices to provide supporting functionality from snippet extraction to calibration with a single, unified API. The platform is self-optimizing; it automatically fine-tunes the neural network hyperparameters and the vector matching sub-systems to optimally configure the tradeoffs between precision, recall, throughput, and latency for each indexed data set.

The Vectara platform operates across languages to deliver a search experience that is truly language agnostic. Search in one language and get the correct result, even if it’s in another language. The Vectara neural-search platform does that by mapping the input query from language space to a meaning space, then it matches meaning-to-meaning to return accurate results, irrespective of language. Most solutions on the market today translate all content and queries to English first to emulate this capability, but that has a whole slew of drawbacks including cost, speed and quality of results. Vectara supports cross-language search for the most widely-used written languages on the planet, including Chinese, Arabic, Spanish, German, Portuguese, Japanese, Korean, French, Italian, Dutch, Polish, Thai, Turkish, Russian, and English.

Try Vectara today

As an API-first company, we know the value of integrating a product into your stack to discover how it works. It’s why we’re launching Vectara with a generous free tier that offers any user up to 15,000 queries per month. For more intensive workloads, our Scale tier provides a complete search pipeline with deeper enterprise features capable of running on the largest websites in the world.

We’re thrilled to begin this journey and spread the power of neural search to organizations of any size. Get started today and connect with us.

• • •

March 5, 2013

What does a CTO do?

Posted in Category: Work — Amr Awadallah @ 9:53 pm | link | | comments (2)

I frequently get asked this question, and it is a question that I sought out to find an answer for a little less than a year ago (when I became full time CTO at Cloudera). I did a bunch of online research and talked to a number of enterprise software CTOs whom I look up to (listed at end). I summarize my findings by answering four core questions: (1) what is the CTO Mission? (2) how to measure the CTO? (3) How should the CTO split his/her time internally vs externally? then finally (4) what are the specific responsibilities of the CTO towards the main constituents inside the organization?

Note that this role description is focused on pure CTOs, as opposed to CTOs whom still own the operational aspects of the engineering organization, it is also biased towards enterprise CTOs vs other domains.

1. What is the Mission of the CTO?

The Mission for the CTO can be summarized across three axes:

A) Fiduciary for Long-Term Technical Strategy:

  • The CTO must effectively hold, develop, articulate, and continually evolve the company’s strategic technical direction.
  • The CTO is responsible for making sure the company continues to have the best technology offering in a dynamically-evolving highly-competitive space.
  • The CTO brings the proper balance between business and technology strategy by effectively connecting the outside world with the inside through distilling information into the key trends that indicate where the company needs to go next, thus aligning the company’s technology strategy with its business strategy.

B) Technical Evangelist:

  • The CTO must inspire/excite people internally around the long-term vision of the company, as well as convince outsiders that this is where the world is going to be, and that his/her company is the best choice to take them there.
  • The CTO must speak with authority about the market needs, must be credible with customers, and must be able to articulate the business value and ROI to a variety of audiences.

C) Spiritual Leader for the engineers and maintainer of the technical culture:

  • The CTO must rally the engineering organization towards the long-term technical goals of the company.
  • The CTO must be able to inspire new engineers to join the engineering organization and must help in the sourcing/identification of such talent.
  • The CTO must help set and maintain the technical culture to make sure the company can continue to retain and attract top technical talent.

2. How to measure the CTO performance?

I could think of three metrics which I outline below:

A) Tech Strategy Alignment:

This is a binary metric, if the company misses a key tech trend then counts against. Also, if there is a significant misalignment between the business and technology strategies then counts against. Towards that end, I highlight this quote which is attributed to Greg Papadopolous [x-CTO of Sun]:

“The CFO is not responsible for making revenue every quarter, but if there is a big surprise, fire her. Similarly, the CTO isn’t responsible for delivering products every quarter, but if he misses a key technical inflection point (e.g. the internet), fire him.”

B) Health of Engineering Culture:

This can be measured by doing a survey of the technical organization to ensure that they are productive and happy at same time. A lagging indicator is significant attrition of top engineering talent, or inability to attract new top talent to the company.

C) Internal CSAT:

In many ways the CTO, similar to the CFO, is a service center to all the other organizations in the company (more on these responsibilities later). Thus an appropriate metric is a weighted customer satisfaction score across these organizations (aka CSAT). If these internal organizations don’t feel they are getting value add out of the CTO, then it is time to fire him.

3. How should the CTO split his/her time internally vs externally?

There is no perfect answer for this one, but the right answer is a bit of both. The CTO will not be able to do a good job inside without having knowledge of the outside, and vice versa. The mix of external/internal is ultimately something that should come out of how the CTO can be of most help to the business. Also there’s travel/family to consider, you have to work sustainably, so find your own sweet spot. The answer will also change as a function of what stage the company is at. With that background this is how I split my time today:

A) External: 70% of time

  • Sales/Customers 35%
  • Marketing/Evangelism/Analysts 20%
  • Business Development/Partners 15%

B) Internal: 30% of time

  • Digest the info collected externally then funnel that knowledge to the appropriate internal teams.
  • Continuously connect engineering, product & business to ensure that the vision articulated publicly is where the company indeed is going, and that the technology roadmap aligns with that vision (especially for new initiatives).
  • Support the marketing team with collateral/white-papers that articulate the company vision.
  • Stay on top of engineering culture, processes, and retention.
  • Shepard for the Intellectual Property portfolio.

4. So what are the CTO’s specific responsibilities for each organization within the company?

 There are five main constituents that the CTO serves: (1) CEO/Strategy, (2) Engineering/Product, (3) Sales, (4) Business Development, and (5) Marketing. Below I list some of the key responsibilities towards each of these constituents.

A) CEO/Strategy

  • Predict and stay ahead of any technical inflection points that might significantly affect the company.
  • Advise the CEO (and CFO/COO) on the long-term technical strategic direction of the company and where to, or to not, make large strategic technical bets.
  • Provide the CEO with different “options” on the technical direction of the company and provide sufficient information for deciding what is the best option to take at any given time.
  • Be a sounding board for the CEO by providing a neutral view which puts the company’s long-term interests above all else. The CTO, similar to the CFO, is in a unique position to do this effectively since he/she doesn’t directly own significant business resources per se.

B) Engineering/Product:

  • Though the CTO isn’t responsible for the day-2-day tactical delivery, he should work closely with the VPs of Products & Engineering to make sure that the overall development direction is well aligned with the company’s strategic technical vision.
  • The CTO should influence the prioritization of resources across large strategic technical bets, but should not be involved in the daily execution once the priorities are set. In fact, it is important to let go entirely of obsessing over the day-2-day execution (which can be hard to let go of), but that is essential so you can focus on the long-term strategic thinking without being encumbered by the immediate constraints.
  • One of the key challenges for the CTO is how to champion people, projects and ideas without stepping on the toes of Product/Engineering management. This is why it is necessary that you seek training to become a strong influencer, and be very selective regarding which ideas you plan to focus your energy on (pick your battles).
  • You need to be a sounding board for the VP of Engineering, allow him to vent, and help him with brainstorming about the different challenges that the development team is facing. In many ways, the CTO and the VP of Engineering need to be bonded at the hip, it is a very critical relationship, so spend significant time on this.
  • Help with the recruiting/retention efforts (this includes ties to Academia).
  • Continuously optimize across the whole organization to avoid any duplication of effort and encourage cross-department fertilization.
  • Ensure alignment of the greater technical organization and, when necessary, arbitrate techno-centric turf scraps, architecture conflicts, etc.
  • Foster innovation by organizing periodic hackathons and by being a steward of innovation in the early stages.
  • Serve as master architect across product lines. I don’t do much of this at Cloudera though, we have a number of truly amazing engineers who are way better architects than I am, so best I can do is to simply get out of their way! (I only get involved if there is a strong conflict, which luckily is rare at Cloudera).

C) Sales:

  • Partner with the sales organization to effectively close customer accounts.
  • Partner with the sales organization to effectively close customer accounts.
  • Partner with the sales organization to effectively close customer accounts.
    (3x replication intended for emphasis)
  • Support the sales organization by building strong relationships with peers at strategic customer accounts.
  • Participate in customer meetings to articulate the long term technology road map hence strengthening the customer’s confidence that our company is leading the industry.
  • Speak with authority about the market; listen to customer needs; quickly understand their issues; and give good advice on the company’s products to the customers.
  • This is very important: know when to say “I don’t know”, then connect the customer with the proper resource whom can go deeper on the question being asked. A good CTO should never ever make shit up on the fly 🙂

D) Business Development and Partnerships:

  • Support the business development team in strategic partner engagements and maintain good relationships with peers at these organizations.
  • Provide technical due diligence of partner technologies and acquisition targets to make sure they properly fit with the company’s platforms, offerings, and culture.
  • The CTO should keep track of all the technology startups in the same space, and have them stack ranked based on what he/she can glean about their prospects. The CTO should have clear thoughts about: possible acquisition targets, what expertise is the company missing? which companies are doing the best work across all of the ancillary areas? which companies have the best technical teams? what could competitors buy that would hurt the company? etc. I rely significantly on my cofounder (Jeff Hammerbacher) for this part.
  • Represent the company at select technical associations to re-enforce the company’s presence and voice with partners.
  • Predict if a new partner technology would have significant impact on the long-term technological roadmap for the company (e.g. a new type of storage or compute device).
  • Predict long-term competitive trends due to the constant shifts in the market.

E) Marketing:

  • Serve as public face of technology for the company.
  • Evangelize the company vision and technical direction through conferences, speaking engagements, and press/media/analyst activities.
  • Maintain good relationships with designated key industry analysts.
  • Support the marketing team in building a large active community around the company’s products (meetups, hackathons, industry conferences, etc).
  • Social engagement marketing through twitter, blog posts, articles/whitepapers, etc.

In summary, a great CTO can be invaluable for their organization as he/she has the luxury to step back and see the forest through the trees, and the ability to be the voice of the customer and of the employee. Whether you are a budding CTO, or a CTO veteran, I really hope you find this article useful. Please don’t hesitate to let me know if you disagree with any of the concepts presented here, or if there are major areas of responsibility that you think I overlooked.


— amr

PS: I would like to acknowledge the following folks/references:

PPS: A good CTO must also know how to be an active participant in the VIM 🙂

• • •

November 15, 2011

The Geometry of Startup Valuations.

Posted in Category: Tech,Work — Amr Awadallah @ 10:40 pm | link | | comments (1)

I frequently find that I have to explain to candidates joining Cloudera (and existing employees) how not to be overly concerned about dilution of their percent ownership when a company is experiencing strong growth.

Towards that end I created the chart below to illustrate how percent ownership is just a small part of the total equation. The core concept is that the “paper” valuation of your stock options is equal to twice the area of your sector under the pie, and the area of your sector is proportional to your angle (i.e. your percent ownership) multiplied by the radius squared (where radius is the stock price). Hence if a startup raises a new round of funding at a large radius then that quickly out weighs any shrinkage in the angle of ownership due to the to the squaring of the stock price. I am not saying that you shouldn’t worry about dilution at all, you should obviously try to get the best deal and keep dilution low, but don’t be overly obsessed by dilution, it is just one factor in the grand scheme of things.

Fortunate startups that demonstrate repeatability/efficiency of their business model raise additional funds at a large radius so that they can fuel further growth thus making the pie larger for everyone. On the other hand startups that get overly concerned by dilution and avoid raising money (depsite being in a growth market) can fall in a penny-wise pound-foolish trap where they end up being a zombie company that goes nowhere (I know many such startups, and by the time they realize their mistake it is usually too late).


— amr

PS: For stock options, the “paper” value is (area of sector) minus (cost of sector), where cost of sector is angle * strike-price^2.

• • •
Next Page »
• 29 queries. 0.392 seconds.