Startups

Can your startup support a research-based workflow?

Comment

Artificial Intelligence Brain
Image Credits: Andriy Onufriyenko (opens in a new window) / Getty Images

João Graça

Contributor

João Graça is CTO and co-founder of Unbabel, an AI-powered language operations platform that enables any agent to communicate in any language.

The President’s Council of Advisors on Science and Technology predicts that U.S. companies will spend upward of $100 billion on AI R&D per year by 2025. Much of this spending today is done by six tech companies — Microsoft, Google, Amazon, IBM, Facebook and Apple, according to a recent study from CSET at Georgetown University. But what if you’re a startup whose product relies on AI at its core?

Can early-stage companies support a research-based workflow? At a startup or scaleup, the focus is often more on concrete product development than research. For obvious reasons, companies want to make things that matter to their customers, investors and stakeholders. Ideally, there’s a way to do both.

Before investing in staffing an AI research lab, consider this advice to determine whether you’re ready to get started.

Compile the right research team

Assuming it’s your organization’s priority to do innovative AI research, the first step is to hire one or two researchers. At Unbabel, we did this early by hiring Ph.D.s and getting started quickly with research for a product that hadn’t been developed yet. Some researchers will build from scratch and others will take your data and try to find a pre-existing model that fits your needs.

From there, you’ll need to hire research engineers or machine learning operations professionals. Research is only a small part of using AI in production. Research engineers will then release your research into production, monitor your model’s results and refine the model if it stops predicting well (or otherwise is not operating as planned). Often they’ll use automation to simplify monitoring and deployment procedures as opposed to doing everything manually.

None of this falls within the scope of a research scientist — they’re most used to working with the data sets and models in training. That said, researchers and engineers will need to work together in a continuous feedback loop to refine and retrain models based on actual performance in inference.

Choose the problems you want to solve

The CSET research cited above shows that 85% of AI labs in North America and Europe do some form of basic AI research, and less than 15% focus on development. The rest of the world is different: A majority of labs in other countries, such as India and Israel, focus on development.

What’s the difference? While Google’s X division may have the capital to focus on moonshots, most startups can only invest in innovation that provides them a competitive advantage or improves their product.

Applied research is often used to solve a specific problem or to aid in the development of a product. This type of research often uses existing, open-source machine learning models with a company’s data sets. This practice may be applied to cut costs, increase efficiencies, measure the success of a product or develop features. Applied research becomes easier with APIs — you can effectively take what comes out of academia, apply it to your product and measure its success based on business KPIs.

Another form of development-focused research is called fundamental research. This is done when you need to invent something for a problem that hasn’t been solved yet or build a data set that doesn’t exist. Our research team focused on solving existing quality estimation problems by inventing a completely new metric for machine translation quality. We open-sourced this development to give back to the community and prevent others from having to reinvent the wheel.

While this effort wasn’t tied to specific product features, it has helped both our team and our customers understand if our product is working as intended and how it can improve. Having a clearly defined business benefit helps with both executive- and board-level alignment.

Use research as a talent funnel

Research shows that demand for AI is back up after a temporary COVID-19 slowdown, accompanied by a major talent shortage — nearly 40% of companies cite a lack of technical expertise as a top reason for not deploying AI technologies.

If you’re starting an AI research lab, it can be helpful to recruit someone who has a dual role within an AI-focused academic institution and within the company. It also helps to be close to a good academic center for access to talent. For example, if you’re in proximity to a major university, you can work with master’s and Ph.D. students and offer summer internships that could convert into full-time opportunities. This opens up access to talent within the university and conversely helps give access to careers students may not have considered.

This practice is common among some of the world’s leading corporate AI research labs. For example, Facebook recruited top AI researcher Yann LeCun eight years ago, and he remains both VP and chief AI scientist at the company and a professor at New York University. Collaborations between the company and the university are common, including a recent development to improve the speed of MRI scans.

Having this type of relationship can help build better bridges and avoid the perception of a “brain drain” of university talent. Remember that your research should ultimately support the company, the university, and its students. As seen in the Facebook example above, opportunities to do joint research projects are immense. For example, we are working on a multilingual conversational agent project with Carnegie Mellon University in Portugal and other academic partners. This type of collaboration benefits everyone involved.

Set expectations on the investment and time to ROI

Most startups and scaleups are used to fast-paced resourcing and R&D cycles designed specifically for product development. However, there is no rushing academic-style research. It’s a rigorous, methodical process that takes time to be successful. Does your culture support the idea of bringing long-term research to reach state of-the-art outcomes? Or are your goals too aggressive?

It can help to lay out the case for a research function to the board early in the process. If the leadership team isn’t fully convinced of the value, you can start small by hiring only one researcher and trying to outsource some of the research to a community (there are existing communities like Kaggle, or in our case, there’s a robust natural language processing community that supports projects with universities).

If you’re working with a university, you can apply for a joint research grant, support a master’s or Ph.D. student, or work on an ongoing project over the course of a year that pairs students with employees at the company.

Investmentwise, an AI lab should be considered in the same category as R&D. It can be easier to get the leadership team or board to consider what they’re willing to invest if you can justify whether your project worked and had an impact on the organization. Some companies are more prone to research because of their DNA, but others will face more of an uphill battle explaining why they’re proposing research and aligning expectations with reality.

Start by really understanding your product as well as which KPIs would be changed if a piece of research were to work. You can use that process to prioritize different research. To prove your concepts, make sure to create datasets that are good proxies to the product KPIs at hand. That way, if anyone asks how to map the value of research back to a product KPI, you’ll have all of the documentation in place to do so.

The power of an AI lab

Setting up and sustaining an AI lab can be costly and can easily become a distraction. For example, one of the biggest cost centers is computing power (such as on-premise or cloud-based GPUs). Anticipating these costs as well as seeking ongoing alignment between research and product is key.

Research has to explicitly include a roadmap and process for incorporating research outcomes into the product. A research arm will fail if expectations are not aligned between the product team and the lab.

With the proper alignment and commitment, you will attract top researchers who will feel fulfilled in their career path, and you can make major advancements you’d never expect from developing a product alone.

These simple metrics will tell you if your startup is ready to scale

More TechCrunch

The Series C funding, which brings its total raise to around $95 million, will go toward mass production of the startup’s inaugural products

AI chip startup DEEPX secures $80M Series C at a $529M valuation 

A dust-up between Evolve Bank & Trust, Mercury and Synapse has led TabaPay to abandon its acquisition plans of troubled banking-as-a-service startup Synapse.

Infighting among fintech players has caused TabaPay to ‘pull out’ from buying bankrupt Synapse

The problem is not the media, but the message.

Apple’s ‘Crush’ ad is disgusting

The Twitter for Android client was “a demo app that Google had created and gave to us,” says Particle co-founder and ex-Twitter employee Sara Beykpour.

Google built some of the first social apps for Android, including Twitter and others

WhatsApp is updating its mobile apps for a fresh and more streamlined look, while also introducing a new “darker dark mode,” the company announced on Thursday. The messaging app says…

WhatsApp’s latest update streamlines navigation and adds a ‘darker dark mode’

Plinky lets you solve the problem of saving and organizing links from anywhere with a focus on simplicity and customization.

Plinky is an app for you to collect and organize links easily

The keynote kicks off at 10 a.m. PT on Tuesday and will offer glimpses into the latest versions of Android, Wear OS and Android TV.

Google I/O 2024: How to watch

For cancer patients, medicines administered in clinical trials can help save or extend lives. But despite thousands of trials in the United States each year, only 3% to 5% of…

Triomics raises $15M Series A to automate cancer clinical trials matching

Welcome back to TechCrunch Mobility — your central hub for news and insights on the future of transportation. Sign up here for free — just click TechCrunch Mobility! Tap, tap.…

Tesla drives Luminar lidar sales and Motional pauses robotaxi plans

The newly announced “Public Content Policy” will now join Reddit’s existing privacy policy and content policy to guide how Reddit’s data is being accessed and used by commercial entities and…

Reddit locks down its public data in new content policy, says use now requires a contract

Eva Ho plans to step away from her position as general partner at Fika Ventures, the Los Angeles-based seed firm she co-founded in 2016. Fika told LPs of Ho’s intention…

Fika Ventures co-founder Eva Ho will step back from the firm after its current fund is deployed

In a post on Werner Vogels’ personal blog, he details Distill, an open-source app he built to transcribe and summarize conference calls.

Amazon’s CTO built a meeting-summarizing app for some reason

Paris-based Mistral AI, a startup working on open source large language models — the building block for generative AI services — has been raising money at a $6 billion valuation,…

Sources: Mistral AI raising at a $6B valuation, SoftBank ‘not in’ but DST is

You can expect plenty of AI, but probably not a lot of hardware.

Google I/O 2024: What to expect

Dating apps and other social friend-finders are being put on notice: Dating app giant Bumble is looking to make more acquisitions.

Bumble says it’s looking to M&A to drive growth

When Class founder Michael Chasen was in college, he and a buddy came up with the idea for Blackboard, an online classroom organizational tool. His original company was acquired for…

Blackboard founder transforms Zoom add-on designed for teachers into business tool

Groww, an Indian investment app, has become one of the first startups from the country to shift its domicile back home.

Groww joins the first wave of Indian startups moving domiciles back home from US

Technology giant Dell notified customers on Thursday that it experienced a data breach involving customers’ names and physical addresses. In an email seen by TechCrunch and shared by several people…

Dell discloses data breach of customers’ physical addresses

Featured Article

Fairgen ‘boosts’ survey results using synthetic data and AI-generated responses

The Israeli startup has raised $5.5M for its platform that uses “statistical AI” to generate synthetic data that it says is as good as the real thing.

16 hours ago
Fairgen ‘boosts’ survey results using synthetic data and AI-generated responses

Hydrow, the at-home rowing machine maker, announced Thursday that it has acquired a majority stake in Speede Fitness, the company behind the AI-enabled strength training machine. The rowing startup also…

Rowing startup Hydrow acquires a majority stake in Speede Fitness as their CEO steps down

Call centers are embracing automation. There’s debate as to whether that’s a good thing, but it’s happening — and quite possibly accelerating. According to research firm TechSci Research, the global…

Retell AI lets companies build ‘voice agents’ to answer phone calls

TikTok is starting to automatically label AI-generated content that was made on other platforms, the company announced on Thursday. With this change, if a creator posts content on TikTok that…

TikTok will automatically label AI-generated content created on platforms like DALL·E 3

India’s mobile payments regulator is likely to extend the deadline for imposing market share caps on the popular UPI (unified payments interface) payments rail by one to two years, sources…

India likely to delay UPI market caps in win for PhonePe-Google Pay duopoly

Line Man Wongnai, an on-demand food delivery service in Thailand, is considering an initial public offering on a Thai exchange or the U.S. in 2025.

Thai food delivery app Line Man Wongnai weighs IPO in Thailand, US in 2025

Ever wonder why conversational AI like ChatGPT says “Sorry, I can’t do that” or some other polite refusal? OpenAI is offering a limited look at the reasoning behind its own…

OpenAI offers a peek behind the curtain of its AI’s secret instructions

The federal government agency responsible for granting patents and trademarks is alerting thousands of filers whose private addresses were exposed following a second data spill in as many years. The…

US Patent and Trademark Office confirms another leak of filers’ address data

As part of an investigation into people involved in the pro-independence movement in Catalonia, the Spanish police obtained information from the encrypted services Wire and Proton, which helped the authorities…

Encrypted services Apple, Proton and Wire helped Spanish police identify activist

Match Group, the company that owns several dating apps, including Tinder and Hinge, released its first-quarter earnings report on Tuesday, which shows that Tinder’s paying user base has decreased for…

Match looks to Hinge as Tinder fails

Private social networking is making a comeback. Gratitude Plus, a startup that aims to shift social media in a more positive direction, is expanding its wellness-focused, personal reflections journal to…

Gratitude Plus makes social networking positive, private and personal

With venture totals slipping year-over-year in key markets like the United States, and concern that venture firms themselves are struggling to raise more capital, founders might be worried. After all,…

Can AI help founders fundraise more quickly and easily?