Byline: Jason Phillips
Anat Heilper may not be the loudest voice in the room, but she is building what everyone else is still trying to comprehend. She is not just a decorated Intel architect; she has spent her life quietly and deliberately wrestling with the complexities of machines, systems, and expectations. Anat’s contributions to AI include three patents on compiler optimization for AI accelerators and receiving Intel’s highest award, the IAA award, for her groundbreaking work. Her superpower lies in her comfort with discomfort and her ability to perform well in one of the tech industry’s most competitive arenas.
With her character and inquisitive mind, Anat is able to solve complex issues quietly without paying attention to her achievements. Her work stands for itself, and this approach allows her to solve problems quietly, even in environments where being loud is the usual method.
From Not Knowing to Mastering
Anat did not enter her chosen field with a five-year plan or any sense of long-term security. She followed her intuition and pursued things that challenged her the most. Anat’s openness to the unknown became her defining edge, where he had to learn more from scratch. Her intellectual curiosity sparked in 2014 when AI became a profound area of interest. Since then, she has played a pivotal role in shaping the development of Intel’s NNP-I and Gaudi-2 AI accelerators, revolutionizing inference processing for data centers. AI entered her life as more than just the latest trend; it ignited a profound intellectual spark in 2014 that has carried her purpose ever since.
Leading by Understanding, not Authority
Anat has quietly earned authority through her depth and knowledge in a culture known for its obsession with titles and visibility. Her leadership on Intel’s NNP-I chip involved overseeing the integration of over 80 engineers to deliver cutting-edge AI solutions, ensuring tight coordination between hardware and software for maximum performance. The hardware/software co-design approach enabled NNP-I to achieve best-in-class performance and power efficiency, reaching 4.8 TOPs/W for major data center inference workloads. It was less about commanding and more about integrating vision. Anat’s leadership ability is rooted in her view of systems as living organisms, where trust is built through technical expertise, not hierarchy.
Curiosity Overconfidence
Where others merely feign certainty, Anat leads the way with questions that many do not dare to ask. Her humility masks a rigorously trained mind, allowing her to spot the weak seams in massive architectures. Anat’s impostor syndrome sharpened her intuitive skills and knowledge, boosting her leadership ability. Her work with the NNP-I and Gaudi-2 chips exemplifies this approach, optimizing AI performance by creating software-hardware integration that outperforms competitors in the cost-to-performance ratio.
Using the Gender Gap as an Engine, not a Barrier
Anat has had to overcome more than a trending wave of AI. She is often the only female in high-level AI meetings, but due to her innate leadership qualities and inquisitive nature, this is just background noise to her. Even among the noise, Anat’s presence matters, and she humbly exudes the confidence and strength to get the job done. Her impact is clear: for BERT Phase-2 training throughput, Gaudi-2 bested the 80GB A100 by a factor of 2.8×, a testament to the effectiveness of her architectural approach. Her approach proves that technical excellence doesn’t have a one-size-fits-all template but instead requires the right questions and confidence to elicit the best answers for the project.
Future Mastery
Now embedded in Intel’s CTO office, Anat is doing what she has always done best: entering the next domain and shaping the foundations before most know the problem. Her contributions continue to push the boundaries of AI hardware and software integration, ensuring that Intel remains at the forefront of innovation.
Anat stands apart because she does not make bold declarations or have an ego-driven vision. She also cannot predict the future. Instead, Anat quietly codes it into being, using her innate talents and inquisitive nature to get to the bottom of things, no matter who is in the room.
Her quiet confidence is built on her fascination with learning how things operate to help build bigger and better models. And, when the next wave of AI arises, there is more than a good chance that Anat has already solved the complex issues with her modernized approach.
