Month: September 2025

• Sooraj Kumar landed an analyst role at Amazon without a referral after a yearlong job search.
• He shared how he responded to a rescinded job offer and what Amazon’s “loop” interviews were like.
• He said he doesn’t think referrals help candidates as much as they used to.

By the time Sooraj Kumar landed an interview at Amazon, he’d already sat through more than two dozen interviews during his yearlong job search. While none resulted in an offer, he said those experiences helped prepare him for the tech giant’s notoriously challenging “loop” interview process.

Kumar’s job search journey began in January 2022, when he moved to the US from Pakistan to pursue a master’s degree in business analytics at DePaul University. He started looking for full-time employment in November 2022, but said the job market was challenging, in part because he wasn’t yet authorized to work in the US — and some employers didn’t want to wait for his work authorization to be processed or commit to sponsoring a visa in the future.

In June 2023, Kumar graduated from DePaul and began interviewing for a business analyst position at an investment firm. After four interviews, he said he received a verbal offer in August, pending a background check. That’s when things went awry.

Based on his research and conversations with friends who’d gone through the process, Kumar said he expected the background check to take no more than 15 days — but day 20 passed, then day 25, with no update. The company doing the background check told him the delay was tied to his lack of prior work experience in the US, which complicated the process.

In September 2023, around the 25-day mark, Kumar said the company told him that, due to business needs, they couldn’t wait any longer for his background check to come through. His offer had been rescinded.

Kumar’s background check cleared after about 35 days. He reached out to the company to share the update — but by then, it was too late.

“That last-minute decision from the company shocked me,” said the 28-year-old, who lives in the Seattle area.

Kumar is among the tech professionals who have struggled to find work in recent years. Since the start of 2025, global tech companies — including Microsoft, Google, and Amazon — have collectively laid off more than 80,000 employees, according to Layoffs.fyi. Tech job postings on Indeed as of July were 36% below early-2002 levels.

This hiring slowdown has particularly affected recent college graduates trying to enter the workforce. That includes the nearly 200,000 foreign students who received F-1 Optional Practical Training (OPT) work authorization last year — and whose visa status depends on staying employed.

Bouncing back from a rescinded job offer

While waiting for his background check to clear, Kumar started an unpaid business analyst internship to pause the 90-day unemployment clock allowed under his OPT visa — helping him keep his visa status in good standing.

But the bad news was that he still didn’t have a full-time, paying job — and he’d lost momentum in his search. Kumar said he’d been applying to around 20 jobs a day for months, but paused once he received the verbal offer.

After taking a couple of weeks to process what happened and plan his next steps, Kumar said he started waking up at 6 a.m. — a couple of hours before his internship began — to apply for jobs. He’d then apply for a few more hours after work, aiming to submit 50 applications a day.

“I refocused, got even more serious about my search, and put in extra effort,” he said.

One of the roles he applied for that September was a business analyst position at Amazon. He submitted the application through the company’s website, without a referral. Kumar said he barely even remembered applying for the role when, in November, Amazon reached out to schedule an interview.

How he prepared for Amazon’s loop interview day

After a one-hour phone interview with someone from the hiring team, Kumar was invited to Amazon’s “loop,” a daylong series of interviews — in his case, five, all of which were conducted virtually. Aside from a 30-minute lunch break, he said the interviews were separated by only about five minutes.

Kumar’s loop included three interviews in the morning, from roughly 9 a.m. to noon, followed by a lunch break and then two more interviews in the afternoon. He said he spoke with members of the hiring team, people from teams he’d collaborate with, and one person from HR. Around Christmas of 2023, Kumar said he received and accepted an offer from Amazon, which came with a six-figure compensation package.

Looking back on the interview process, Kumar said it was important to be familiar with Amazon’s leadership principles — including “customer obsession,” “ownership,” and “invent and simplify” — and to prepare professional stories that align with these values.

He said that one effective, Amazon-approved way to discuss past work experiences during interviews is the STAR method: situation, task, action, and result. The format helped him clearly explain his impact, and he believes it also made it easier for interviewers to take notes and remember his responses.

While his yearlong job search was grueling, Kumar believes the experience he gained interviewing for other roles helped him excel in Amazon’s process.

“If I didn’t have all the previous interview experience, I might have never ended up at Amazon,” he said.”

Top job-search takeaways

In addition to the value of interview preparation, Kumar said he had two other takeaways from his job search. First, while the quality of an application matters, he felt that volume was just as important — if not more so. No matter how strong his application was, he said the odds often felt stacked against him, so his best strategy was to give himself as many chances as possible.

His second takeaway: Referrals weren’t particularly effective. He said most of the interviews he landed came from cold applications, while the ones he submitted with referrals typically went nowhere.

Kumar suspects referrals have become too common at large companies, adding that his own LinkedIn inbox started filling up with people asking him for one after he joined Amazon. Unless the person referring you is on the hiring team or closely connected to it, he said, he wouldn’t count on a referral having much impact.

“Referrals are very mainstream now,” he said. “I don’t think that many referrals carry much weight.”

• Phil Shawe is the cofounder of TransPerfect, a language and AI solutions company launched in 1992.
• TransPerfect has acquired nine companies since January, including Blu Digital Group.
• Shawe told Business Insider what he looks for in merger partners.

This is an as-told-to essay based on a conversation with Phil Shawe, the cofounder and CEO of TransPerfect. The company, founded in 1992, has acquired nine companies since January: H2A, Technicolor Games, Apostrophe Group, Bear Down Studios, Blu Digital Group, Speech, MPC, The Mill Studios, and Unbable. The essay is edited for length and clarity.

Acquisitions have been near and dear to our strategy. TransPerfect tries to find management teams that are like-minded or have really interesting technology that fits well within our group of technologies.

We don’t approach mergers and acquisitions like a private equity firm might, which could involve buying companies, making positions redundant, and laying people off to make the unit more profitable.

If we don’t feel like the people are a good mix, we probably won’t do the deal. Here’s what I look for in an investment.

Entrepreneurs

The No. 1 thing TransPerfect is looking for is entrepreneurial management.

One reason I love working with entrepreneurs so much is that they’re able to make excellent decisions fast. They’re also willing to change course if they make a decision and later decide it wasn’t correct. I’m looking for entrepreneurs to join TransPerfect and grow their business within a division of the firm.

Our pitch to them is, “You didn’t get into this business to do back-office accounting. Let us handle all that. You go out and do what you love to do, which is take care of customers in your field, build great technology, and sell it.”

Founders or senior executives who stay post-acquisition

We don’t love it when the founders or senior executives leave. We would rather have them join our management team.

We go into acquisitions striving to keep the vast majority of employees and repurpose them. There are people who end up leaving, but we generally try to keep the team together because that’s what we’re acquiring — the human talent and then the technology.

We don’t even call our deals acquisitions. We call everybody a merger partner.

Back in the old days, we would only consider founder-led businesses where the management stayed on. Now, after we have grown and some of the deals have grown, we are willing to consider private equity or other deals where management has left.

Companies that already use TransPerfect’s technology

If someone is running a competitor tool and we acquire them, one of the minimum things we would expect is that they change and use our tools.

Their day-to-day business won’t change, but we’re going to need to change their accounting system to be on our global system and change their language technology platform. We don’t want to use and pay for our direct competitors technology when we have a technology of our own. Our technologies are No. 1 in the market.

A sellable product

If we feel like our existing salesforce can’t sell the product no matter what, even with training or a specialized salesperson, then that might not be as attractive.

Some people acquire businesses just for revenue or logos, but we rarely pursue acquisitions just to generate revenue growth. That ends up happening for us, but it’s not a goal for us because we’re a private company. We’re not trying to show growth to anyone. We’re trying to run a great business for our customers.

Companies that diversify TransPerfect’s holdings

Acquiring another language company might not excite us unless it’s in a new market because we’re always looking to diversify our offerings and be a value-added partner to our customers.

For example, we acquired Blu Digital, a company with amazing technology for distributing film and streaming content. We also acquired MPC, a leading film VFX company that worked on “Dracula.”

Acquiring companies in different time zones also means the sun never sets on our business.

We love the fact that we have production in different time zones in the world, working on projects for customers. If they need something done in a hurry, we can do that without crushing our US staff because we have so much staff internationally.

At the end of the day, though, these are just factors that could weigh one way or the other. To be a great company, we need entrepreneurial management and preferably cool technology that solves a problem better or a problem we don’t address in our current technologies.

• Walmart’s Sravana Karnati said he looks for two things in engineer hires: domain expertise and the ability to learn.
• He said he values subject-matter expertise in computer science over specific programming languages.
• The executive also said he’s looking for people who can keep up with technology’s evolution.

Walmart executive Sravana Karnati has over 25 years of leadership experience at companies like Amazon, Disney, and Oracle — and he looks for two key traits when he hires engineers.

“I look for the fundamentals and ability to learn,” the executive vice president of Global Tech platforms told Business Insider.

Karnati said while he’s never hired a candidate for “specific knowledge,” he wants to know they have domain expertise. Anyone in a technical position, including those in project manager roles, need to understand the architecture, dependencies, and risks involved in the work they do, the executive said.

“It is not about whether you know Java, whether you know C++, or some other language,” Karnati said. “What we look for is, does a person understand the fundamentals of computer science?”

Karnati added that the specific skills someone needs depends on the role they’re aiming for. For example, someone in a UX design or engineering role doesn’t necessarily need to know computer science fundamentals in depth, but some basic technical knowledge could be useful. What’s most important is the ability to effectively use tools like Figma and other LLMs to do the work, he said.

Those in software development or engineering roles, though, need a strong foundation in computer science fundamentals, Karnati said. LLM tools can help make those employees more productive, but they can’t serve as a substitute for a deeper understanding of systems, algorithms, and architecture.

Karnati, who played a key role in developing Walmart’s new AI tool called Wibey, said that he wants candidates who know how to write good algorithms and long-term “sustainable code.” (The company says Wibey is a developer-focused tool that helps employees streamline workflows by automating repetitive coding tasks and managing compliance checks.)

While Karnati said he’s looking to hire computer scientists, that doesn’t necessarily mean the people he hires need a computer science degree. Karnati himself has a doctorate degree in chemical engineering but has been a computer scientist for years, he said.

“If you have an industrial-engineering background with some computer-science orientation, you could do really well in technology operations,” Karnati said, adding that operational rigor is needed in those kinds of roles.

Karnati said “the trick” is keeping up with the evolution of technology and continuing to retrain yourself. Those who have successfully worked in the industry for 30 to 40 years have kept up with the business and continued to learn, Karnati said.

That ability to learn quickly and keep up with the pace of the industry, he said, is the second trait he looks for. When he looks to hire someone, Karnati said he wants to see that the person has already demonstrated that trait.

He said candidates can show this ability by participating in different kinds of internships, or through coursework that indicates a “wide palette of learning, rather than a narrow focus.”

Karnati’s comments align with a broader trend among tech leaders, who have pointed to the growing importance of adaptability in the workplace, as AI becomes ubiquitous.

Waze cofounder Uri Levine previously told Business Insider that the “most important learning today is the ability to adapt.” Similarly, Cloudflare CEO Matthew Prince told Business Insider that the cybersecurity giant is focused on recruiting talent with a “broad set of skills.”

• Boston Consulting Group is integrating AI into metrics for performance evaluation.
• BCG says AI adoption has reached 90%, with 50% of employees using it daily.
• BCG leads in custom GPT development, enhancing performance reviews, and client services.

There is no room at Boston Consulting Group for AI skeptics.

About a year ago, the firm began building AI expectations into the benchmarks that shape how consultants are assessed, Alicia Pittman, the global people team chair at BCG, told Business Insider.

“There’s no box on our forms that says, ‘Are you using AI?’ It is an expectation,” she said, but the technology is now central to the core competencies, like problem solving and insight, that drive evaluations and promotions.

“If you’re not, you won’t do well on the competencies—you’ll fall behind your peers on problem solving and insight,” she said.

AI hasn’t changed the type of work BCG expects from employees, Pittman said, but it has raised the bar on quality and efficiency.

“For example, in problem solving, a consultant might use AI to surface insights — but performance is assessed on the judgment they apply to interpret those insights, structure the problem, and deliver solutions that matter for clients,” she said.

Perhaps the most impactful application of AI is in performance management itself. Pittman said one of the firm’s best AI tools helps employees write performance reviews — cutting writing time by 40% while improving quality metrics by 20%.

BCG’s employees are taking up AI faster than expected

Now, almost 90% of the firm’s 33,000 employees use AI, and about half are what BCG calls “habitual users,” or those who use the technology daily. “That’s something we measure because it leads to stickiness and sophistication of use,” Pittman said. The firm had aimed for 50% adoption by year’s end but hit that milestone in May, months ahead of schedule.

BCG isn’t unique in urging employees to adopt AI. Across the industry, consulting firms are retooling their workforces. Accenture, for example, has said it is “exiting” employees it can’t reskill, even as its CEO, Julie Sweet, projects overall head count will grow in the next fiscal year. At McKinsey, over 70% of the firm’s 45,000 employees now use use its chatbot Lilli, McKinsey senior partner Delphine Zurkiya told Business Insider in March.

Progress at BCG has largely been thanks to its AI training program.

As of April, the firm had developed eight or nine internal tools. Deckster, a slideshow editor trained on 800 to 900 templates, helps consultants build and grade presentations; about 40% of associates use it weekly, Scott Wilder, partner and managing director at BCG, previously told Business Insider.

GENE, a GPT-4o-powered chatbot from ElevenLabs, a startup that develops text-to-voice AI technology, features a deliberately robotic voice and is used for brainstorming, podcasts, and live demos.

The firm provides training for employees on how to use its AI tools. It has a dedicated generative AI learning and development team staffed by rotational consultants who teach and codify best practices. It also has a network of 1,200 people in local offices who function like “ground troops,” Wilder said. They provide hands-on training, gather feedback, and push forward on adoption.

Wilder told Business Insider the firm estimates employees reinvest about 70% of the time they save into “higher-value activities” including analysis and communicating insights clearly.

The firm is also tracking a group of about 1,500 “advanced users” — employees at the edge of experimentation with AI. Roughly two-thirds are associates and consultants, BCG’s entry-level ranks, Pittman said.

One of their biggest contributions has been creating custom GPTs, or no-code versions of ChatGPT tailored to specific tasks.

Consultants have developed GPTs to review slide decks, check clarity, anticipate client questions, and enforce BCG formatting guidelines. Many are tested internally before being shared with clients.

BCG is now the highest number of custom GPTs of any OpenAI customer, Pittman said, with five times more employees building them than a year ago.