EVC Computer Science Grand Meeting

On the 7th of May, 2024, Evergreen Valley College became a hub of technological exchange and innovation as the Computer Science Club hosted a significant meeting on campus from 3:40 to 5:40 pm. This event was not confined to the academic community but extended its reach to faculty and locals interested in the evolving landscape of technology.

The meeting commenced with an introduction by Zakaria Kortam, the President of the Computer Science Club, who set the stage for an afternoon of insightful discussions and groundbreaking presentations. The highlight of the event was the presence of two distinguished guest speakers, each bringing unique perspectives from the forefront of technology and entrepreneurship.

Image of Zakaria Kortam (Note: Image taken at a separate meeting event)

Dinesh Maheshwari’s Presentation on Groq’s Innovative LPU

The first guest speaker, Dinesh Maheshwari, is the Chief Technology Officer and Adviser at Groq, an Artificial Intelligence startup currently valued at nearly $400 million. Maheshwari’s talk centered on Groq’s pioneering Language Processing Unit (LPU), a novel technology designed to elevate AI compute efficiency dramatically.

Maheshwari began his presentation by explaining the concept of the LPU. Unlike traditional computing paradigms, the LPU is engineered specifically for generative AI applications, promising to transform AI’s landscape by offering a more efficient, scalable alternative to conventional GPUs. He elaborated on how LPUs, with their “programmable assembly line” architecture, offer deterministic processing that enhances data handling and instruction execution at the clock level—contrasting sharply with the “hub-and-spoke” architecture of GPUs which often becomes a bottleneck in AI tasks.

A significant part of his talk was dedicated to the efficiency and scaling advantages of LPUs. These units minimize external memory reliance and optimize data flow via high-bandwidth “stream registers.” This architecture supports quicker, more cost-effective scaling, which is crucial for managing the ever-increasing size of AI models—a challenge that current GPU-based systems are struggling to keep up with.

Maheshwari provided compelling performance metrics, demonstrating that LPUs outperform GPUs in throughput and energy efficiency. He also discussed the practical applications of LPUs, particularly in inference tasks where they ensure high token throughput per user at lower costs and reduced power consumption.

Concluding his presentation, Maheshwari shared Groq’s vision for LPUs in fueling the AI revolution, emphasizing their potential to make AI more accessible and efficient. He encouraged the audience to engage with this emerging technology, highlighting Groq’s commitment to continued innovation in AI hardware.

Following Maheshwari’s presentation, the audience participated in a Q&A session, probing the future of LPU technology and its comparative prospects against GPUs.

Insights from Greg Tanaka, CEO of Percolata

The second speaker, Greg Tanaka, CEO of Percolata—a Series-A startup leveraging machine learning for enhanced marketing and operations—shared his insights on leadership in the tech industry. Tanaka, who is also a Palo Alto City Council member and a congressional candidate, emphasized the essential skills for leading in the tech sector.

He discussed the stark contrasts between working at large, established companies versus dynamic startups. Tanaka underscored the importance of sales and marketing skills, vital not only for promoting products but also for professionals to market themselves effectively within the AI industry.

After his talk, Tanaka engaged with the students in a Q&A session, where he further elaborated on the critical competencies for aspiring tech leaders and entrepreneurs.

Showcase of Student Robotics Projects

Post-discussions, the meeting shifted focus to practical applications of technology by club members. Three groups showcased their robotics projects, each featuring innovative design and functionality:

• The first group, comprising Nicolas Ojeda, Phuong Nguyen, Niya Justine, and Tristan Lee, presented “Dum-E,” a robot equipped with tank tracks, multiple sensors, and rotating cameras, capable of maze solving, pathfinding, and object relocation, all controlled via a mobile device.
• The second group, “L.A.G.S,” consisting of Vince Ly, Ariana Estrada, Brady Gho, and Sherilyn Huynh, demonstrated a similar tank-tracked robot, emphasizing user-friendly mobile control interfaces.
• The final presentation by “The Best Group” — Khai, Clive, and Emi — featured an autonomous robotic vehicle adept at navigating various terrains autonomously, equipped with cameras and sensors for environmental detection and path optimization, highlighted by a live demonstration of its capabilities, including a facial recognition system.

*Images in order of the groups (Left to right)

Following the robotics presentations, the outgoing administration of the Computer Science Club summarized the club’s achievements over the past academic year. Professors Henry Estrada and Kathy Jensen, advisors of the club, also contributed their insights. The club reported notable numerical achievements, including a membership of 64 official members and the successful completion of seven projects. Throughout the year, the club organized 26 meetings and hosted numerous events that attracted thousands of participants. Additionally, the club’s social media presence on platforms such as Instagram, LinkedIn, and X achieved over 20,000 engagements, highlighting its significant digital footprint and active community involvement.

This grand meeting not only provided a platform for sharing cutting-edge technological advancements but also encouraged students to engage directly with industry leaders and apply their knowledge in practical, creative ways.

The event successfully fostered a spirit of innovation and curiosity among the participants, underscoring the vibrant academic community at Evergreen Valley College.