AI Software Engineer Intern

Intel gets a packaging bump. Intel shares jumped more than 12 per cent after Chipzilla signed the company which sounds like a sneeze Hitachi for advanced packaging work aimed at AI and data centres. Intel announced a partnership with Hitachi to co-develop and produce advanced semiconductor packaging technologies. The deal focuses on high-performance kit for data centres, artificial intelligence systems and high-end computing. Chipzilla has been under pressure from investors and customers after manufacturing delays and rising competition from TSMC and Samsung Electronics. The Hitachi deal gives it another way to talk about progress without waving another delayed node roadmap around. The two companies will work on technologies that stack multiple chips together to improve speed and cut energy use. Hitachi brings materials science and precision manufacturing experience, which Chipzilla clearly thinks it could use. The collaboration will focus on hybrid bonding and other next-generation packaging methods. These techniques allow finer connections between silicon components, boosting bandwidth between processors, memory and accelerators while reducing latency and power draw. Chipzilla has pushed advanced packaging as one of its key weapons against foundry rivals. Its EMIB and Foveros technologies already appear in products including some Xeon processors and the Ponte Vecchio data centre GPU. Working with Hitachi could speed development of more sophisticated stacking methods for future products. The Japanese industrial giant has long supplied critical materials and equipment to the global semiconductor industry. The share jump reflected investor relief after Chipzilla's market value took a kicking for three years. Shares had fallen more than 60 per cent from their 2021 peak before the announcement. The Hitachi pact is part of a wider supply-chain strategy. Chipzilla has been expanding its foundry services business to attract customers that might otherwise hand everything to TSMC. The company has broken ground on new fabs in Ohio, Arizona and Germany as it tries to rebuild manufacturing capacity in the US and Europe. Support from the CHIPS and Science Act has brought billions in grants and loans to offset the eye-watering capital costs. Hitachi sees the deal as a way to expand in advanced packaging. The company already makes specialised materials used in semiconductor assembly and has invested in thermal management and high-density interconnects. Analysts reckon the collaboration could help both companies tackle technical problems slowing progress across the sector. As transistors hit atomic-scale dimensions, raw processing gains from lithography alone are getting harder to find. Chiplet architectures have already shown their worth in graphics processors and server CPUs from AMD and Nvidia. That makes packaging less of a back-room engineering detail and more of a frontline fight. Industry observers expect initial prototypes within 12 to 18 months, followed by possible commercialisation if big customers approve. The technologies could support server processors, AI accelerators, networking chips and scientific computing systems. Chipzilla has picked up design wins for its upcoming Lunar Lake and Arrow Lake consumer processors. The Hitachi partnership gives investors another reason to believe the turnaround story is not entirely held together with tape. Jun 12, 2026

Joan Tafoya, former director at Meta, Intel & Sandia: why swarming every problem slows teams. by Jamie Flinchbaugh on 06-11-26 | / | / | / | Joan Tafoya joins Jamie Flinchbaugh on People Solve Problems for a conversation built around one deceptively simple question: how do you coach problem solving when the person in front of you is difficult to coach? Joan, a former director at Meta, Intel, and Sandia National Laboratories with nearly 38 years of experience, brings a perspective shaped by leading large, globally dispersed teams and by watching problem-solving succeed and fail at scale. Joan opens by describing her own growth as a leader. In her early years as an engineer and manager, she was focused on results and often grew frustrated when people did not solve problems the way she would have. Over time, she recognized that she could not advance until she taught others to think differently, not to think like her, but to think more deliberately about problem-solving itself. She points to a turning point while running a manufacturing line at Intel, where a constant stream of people asking for the next step left her burning out. The shift she made was to narrate her own thinking out loud so others could follow it, and to ask people what they were thinking rather than supplying every answer. That experience leads to one of the central themes of the episode. Joan is honest that her move toward coaching began partly as self-preservation, but she came to see it as something larger. She realized her team did not always share the same picture of success, and that alignment on both the problem and the desired outcome mattered as much as effort. Coaching, in her telling, became the way to build a group of trusted people who could carry the work forward and free her to take on new challenges. Joan also reflects on the challenge of coaching highly credentialed experts. At Sandia National Laboratories, she built a department focused on lean and problem-solving practices in an environment where nearly everyone held a PhD. Rather than pretending to match their technical depth, she earned credibility by listening carefully, reflecting what she heard, and asking sharper questions, especially about the knowledge gaps standing between the team and its next breakthrough. Prioritization is another area Joan explores in depth. Drawing on her time at Meta, she describes a culture that was learning to balance speed with reliability. Her first prioritization question is not whether a problem is important but whether her team is the one best suited to solve it. She warns against the instinct to swarm every visible problem, comparing it to a kindergarten soccer team chasing the ball. She also looks for quick wins that build momentum and morale without draining resources from larger work. On ideation, Joan emphasizes starting with clarity. She often opens sessions with a silent exercise in which everyone writes down the problem statement and what success would look like, which surfaces a wider range of views and avoids groupthink. She values bringing contrarians into the room, while noting that a contrarian works best when given clear expectations about how to contribute rather than simply being allowed to push back. Joan closes with advice for young engineers hoping to lead in fast-paced, demanding environments. She believes people grow most while they are struggling and pushing past what feels comfortable, and she encourages giving yourself grace during those stretches while still choosing to step into the discomfort. It is a fitting end to a conversation that treats problem-solving as something every person can learn and every leader can nurture. You can connect with Joan Tafoya on LinkedIn