Intel Continues to Rehire Veterans: At Some Point They’ll Run Out
by Dr. Ian Cutress on July 7, 2021 4:13 AM EST
News on the wire today is that Intel has rehired 28-year veteran Shlomit Weiss into the position of Senior VP and Co-General Manager of Intel’s Design Engineering Group (DEG), a position recently vacated by Uri Frank who left to head up Google’s SoC development. As reported in Tom’s Hardware and confirmed in her own LinkedIn announcement, Weiss will be working at Intel’s Israel design center alongside Sunil Shenoy and is ‘committed to ensuring that the company continues to lead in developing chips’. Weiss is the latest in an ever-growing list of ‘re-hiring’ Intel veterans, which leads to the problem that at some point Intel will run out of ex-employees to rehire and instead nurture internal talent for those roles.
In her first 28-year stint at Intel, Weiss is reported to have lead the team that developed both Intel Sandy Bridge and Intel Skylake, arguably two of the company’s most important processor families over the last decade: Sandy Bridge reaffirmed Intel’s lead in the market with a new base microarchitecture and continues in its 6+th generation in Comet Lake today, while Skylake has been Intel’s most profitable microarchitecture ever. Weiss also received Intel’s Achievement Award, the company’s highest offer, but is not listed as an Intel Fellow, while CRN reports that Weiss also founded the Intel Israel Women Forum in 2014. Weiss left Intel in September 2017 to join Mellanox/NVIDIA, where she held the role of Senior VP Silicon Engineering and ran the company’s networking chip design group.
In her new role at Intel, Tom’s is reporting that Weiss will lead all of Intel’s consumer chip development and design, while the other Co-GM of Intel DEG Sunil Shenoy will lead the data center design initiatives.
If you’ve been following the news of Intel’s personnel of late, you might start to learn a pattern:
- Dec 20: Intel hires Masooma Bhaiwala (16-year AMD veteran)
- Jan 21: Intel rehires Glenn Hinton (35-year Intel veteran, Senior Fellow)
- Jan 27: Intel rehires Sunil Shenoy (33-year Intel veteran)
- Jan 27: Intel hires Guido Appenzeller (various)
- Feb 15: Intel rehires Pat Gelsinger (30-year Intel veteran)
- Mar 17: Intel rehires Sanjay Natarajan (22-year veteran)
- May 28: Intel hires Ali Ibrahim (13-year AMD veteran, Senior Fellow)
- June 7: Intel hires Hong Hao (13-year Samsung veteran)
- June 8: Intel rehires Stuart Pann (33-year Intel veteran)
- June 8: Intel rehires Bob Brennan (22-year Intel veteran)
- June 8: Intel hires Nick McKeown (27-year Stanford professor)
- June 8: Intel hires Greg Lavender (35-year Sun/Citi/VMWare)
- July 6: Intel rehires Shlomit Weiss (28-year Intel veteran)
Of these named hires (plenty of other people hired below the role of VP), seven are listed as ex-Intel employees being rehired into the company, mostly into engineering-focused positions. These ex-Intel engineers have a long line of accolades at the company, having worked on and built the fundamental technologies that power Intel today. The exact reasons why they left Intel in the first place are varied, with some peers are keen to cite brain drain during CEO Brian Krzanich’s tenure, however it would appear that the promise of working on fundamental next-generation hardware, along with popular CEO Pat Gelsinger, is enough of an allure to get them to return.
It should be noted however that number of engineers that Intel could rehire is limited – going after key personnel critical to Intel’s growth in the last few decades, despite their lists of successful products and accolades, can’t be the be-all and end-all of Intel’s next decade of growth. If we’re strictly adhering to typical retirement ages as well, a number of them will soon be at that level within the next ten years. Intel can’t keep rehiring veteran talent into key positions to get to the next phase in its product evolution – at some level it has to reignite the initial passion from within.
Intel’s key personnel are often home-grown, or what we call ‘lifers’, who spend 20+ years of the company typically straight out of university or college – every rehire on this list fits into this image, especially CEO Pat Gelsinger, and a number of contacts I have within the company are identical. However if Intel is having to rehire those who enabled former glory for the company, one has to wonder exactly what is going on such that talent already within the company isn’t stepping up. At some point these veterans will retire, and Intel will be at a crossroads. In a recent interview with former Intel SVP Jim Keller, he stated that (paraphrased) ‘building a chip design team at a company depends on volume – you hire in if you don’t have the right people, but if you have a team of 1000, then there are people there and it’s a case of finding the right ones’. In a company of 110000 employees, it seems odd that Intel feels it has to rehire to fill those key roles. Some might question if those rehires would have left in the first place if Intel’s brain drain had never occurred, but it poses an interesting question nonetheless.
Source: Tom’s Hardware, CRN
Image: LinkedIn
Related Reading
- Intel Hires a New Technical Focused Chief Strategy Officer
- Intel Hires Fab Veteran, Former GlobalFoundries CTO Dr. Gary Patton
- New Intel CEO Making Waves: Rehiring Retired CPU Architects
- Intel Appoints Pat Gelsinger as New CEO, From Feb 15th
- Intel to Develop Discrete GPUs, Hires Raja Koduri as Chief Architect & Senior VP
Facebook
Twitter
RSS
121 Comments
View All Comments
Dehjomz - Wednesday, July 7, 2021 - link
They have healthy revenues now, but with a resurgent AMD, Apple, Qualcomm and NVidia pushing ARM solutions, it is urgent to develop the war chest NOW. Sometimes the best offense is a good defense, and that’s the strategy they’re going with. Blunt the impact of the competition, with trusted leadership, and when the competition is licking their wounds, then you bring in the next generation of leaders. Let’s see how the strategy plays out in the next 3-5 years.mode_13h - Wednesday, July 7, 2021 - link
> There’s a reason why Intel was producing rather high performance coresSo, the old dogs knew how to build good CPUs at yesterday's scale and with yesterday's tech. It's not obvious they're the right ones to lead Intel's foray into new generations. I'm not being age-ist, here. Maybe they are, but it's not a given.
It just feels too much like replaying the greatest hits. Although there's a lot to be said for Pat having a team under him that he trusts. Maybe that's a lot of what this is about.
Dehjomz - Wednesday, July 7, 2021 - link
This is the lady that led the Skylake team, Intel’s most profitable architecture. Smart to bring her back, especially now that they have EUV working, as well as access to TSMC. Perhaps she’ll go on to lead the development of the next generation of Intel’s most profitable architecture.whatthe123 - Wednesday, July 7, 2021 - link
old dogs built an architecture that's nearly as performant as zen 2 per core, on a worse process with fewer transistors. how much time do you think has passed exactly? It's been less than a year since zen 3.Not to mention that zen's designs were finalized by "the old dogs" like Mike Clark and Jim Keller. Turns out old dogs can learn new tricks.
mode_13h - Wednesday, July 7, 2021 - link
> an architecture that's nearly as performant as zen 2 per core,> on a worse process with fewer transistors.
And that's leakier than a sieve.
And if we're comparing across process nodes, then you should also acknowledge that it's a lot less efficient than Zen2. They kept it in the game by massively juicing the TDPs and cranking the GHz.
> zen's designs were finalized by "the old dogs" like Mike Clark and Jim Keller.
Jim said he's not the father of Zen. But rather than get into exactly what he did do, let's try to get back on track.
I gather you're triggered by my "old dogs" comment, but my real point was about trying to recapture the success of the past by replaying the greatest hits. Except, that's not how it works. In the age of AI chip design and changing HW/SW boundaries, all I'm saying is that I wouldn't presume the old guard is necessarily the best to lead that change. Not to say that some aren't, but it's not obvious that he's really focused on how Intel needs to adapt for the future, rather than just trying to Make Intel Great Again.
whatthe123 - Wednesday, July 7, 2021 - link
It's leaky because their process is leaky. They've been forcing incremental gains by widening and pushing boost for years, has hardly anything to do with the fundamental architecture, as you can see with rocketlake fattening up and still eating absurd amounts of power.And it's ironic that you talk about the "new age" of chip design when all of the market leaders right now have teams that are being overseen by the old guard. Hell article is about Weiss, who went from flipping Mellanox to integrating with Nvidia, and now nvidia basically owns the AI market. Jensen's still leading nvidia with no signs of handing it off. You've got Keller working with Ljubisa Bajic, another old hat from the ATi days. Just because these old people lead the company doesn't mean they are stamping out the ideas of their younger, fresh employees.
whatthe123 - Wednesday, July 7, 2021 - link
Hell Renée James, another person from the 80s era intel, founded ampere computing, which is probably the most successful enterprise ARM cpu developer right now. Man, someone should've let her know what she should've just given up because of her association with intel.mode_13h - Wednesday, July 7, 2021 - link
> ampere computing, which is probably the most successful enterprise ARM cpu developerThey just took IP out of ARM's parts bin and slapped it together to make Altra. There's nothing original in there, as far as we know. Even the power management is basic.
We can't really know if Ampere is any good until they release their in house core and we see how it stacks up against CPUs with the N2 cores.
mode_13h - Wednesday, July 7, 2021 - link
And they weren't even first to market with a Neoverse CPU. Amazon beat them to market by like 10 months and probably shouldered most of the burden of debugging ARM's IP.mode_13h - Wednesday, July 7, 2021 - link
> It's leaky because their process is leaky.No, I mean it leaks information almost faster than Zen can compute it! It's leaky because Intel employed a bunch of optimizations and cut corners where they shouldn't have.
> it's ironic that you talk about the "new age" of chip design when all of the
> market leaders right now have teams that are being overseen by the old guard.
I'm talking about tomorrow and you're still talking about yesterday. My question was whether Pat is putting together the right team to build future generations of chips, which will have to go beyond the existing playbook to stay relevant. I think we won't know for another 2-4 years.
I get that you're triggered by the "old dogs" comment, but try to see past that.