2023 tech interns tackle AI development at the AWS Summit

Over the summer, ambitious college students set down their textbooks to find their industry calling through internships. Fortunately, we found Capital One. Initially, we joined Capital One’s Technology Internship Program (TIP) to gain hands-on workplace experience and expand our industry perspective where leading software innovation meets modern banking and finance. 

We continue to discover is so much more; professional and educational growth opportunities extend far beyond the badge scanners outside our New York office. With our projects acquiring final sign-off and shipping to production, we identified a unique opportunity to attend the Amazon Web Services (AWS) Summit in late July, 2023, hosted in our Empire State backyard.

On top of collecting endless swag (mostly socks) and capitalizing on a free lunch, the AWS Summit was an unforgettable experience fueled by a jam-packed agenda. In the morning, we engaged with engineers from over a hundred pioneering vendors and AWS partners, understanding their products and differentiating factors. Around noon, we listened in on the center-stage keynote, sitting front-row to hear about the company’s product vision, notable innovations and the unveiling of new services. 

Over 115 instructive sessions and workshops were scattered throughout the day, where we became hands-on with AWS services applied towards high-demand use cases. As aspiring software engineers passionate about leveraging next-gen technology for good, the AWS Summit was nothing short of a blast.

Unsurprisingly, AWS mirrored other tech giants in establishing a clear stake in the generative artificial intelligence (AI) movement. 

Powered by recent breakthroughs in Machine Learning (ML) and hardware efficiency, generative AI encompasses extracting insight from multi-modal observations and apply them in highly digestible, actionable form. Now, AWS has officially entered the game. The cloud leader already empowers at-scale software solutions globally with over 200 managed service offerings. Amazon’s new commitment disrupts the software development landscape by seamlessly placing increasingly advanced AI innovations in the hands of platform users—literally just clicks away. The upper bar of software capabilities has shifted beyond return, just as our enthusiasm grows without limit.

Several existing AWS services received functional upgrades, catalyzed by under-the-hood generative AI integration. These upgraded services include: 

• Quicksight Q - EnhancedSmarter business intelligence presentation

• Code Whisperer for Glue Studio - Autocompletion for data engineering

• AWS Healthscribe - Clinical note assistant

In addition to leveraging generative AI within the platform itself, AWS announced a new class of services that facilitate or ease the onboarding of AI practices for customers. Some examples include:

• AWS Bedrock - Foundational models (FMs) as a service

• Agents for AWS Bedrock - Managed FM fine-tuning and integration

• Entity Resolution - Connected record management

• AWS vector engine for Amazon OpenSearch serverless - A step towards native AI search

• EC2 T5 instances - Hardware with generative AI in mind

Like any other technology conference, AWS Summit 2023 NYC aimed to cultivate excitement surrounding its innovations; the cloud giant surely did not disappoint. Each year, AWS expands and improves its cloud offerings to benefit enterprises and developers alike. Walking around the Expo floor, our favorite question to ask vendors was: “How long have you been in business?” It was eye-opening to hear how recently most of these companies were founded. Moving from idea to implementation on AWS has never been easier and the Expo floor was filled with exciting ideas on what to build next. 

Still hunting for coffee in the Expo hall, we asked vendors how they utilize the Cloud for their business-specific needs. We heard from database, security, developer operations (DevOps), and analytics cloud companies. More than ever, the ecosystem created by AWS expands outside the console to drive innovative solutions to market with high velocity and technical accomplishment. That’s a win for all tech-first organizations. It is hard to imagine what the same Expo hall will look like in just a few years or months!

• Amazon maximizes the power of cloud-enabled generative AI by providing flexibility to AWS users and their diverse product needs. Many users and organizations have an existing footprint in the AWS cloud: Storing data, utilizing compute power, centralizing BI, etc. Their commitment to AWS as a core infrastructure platform is rewarded by all-in-one, out-of-the-box services like Bedrock. For these groups, specialized AI solutions are clicks away, leaving behind all the guesswork and advanced engineering demands. 

• AWS offers the individual building blocks necessary to engineer and deliver user-managed AI solutions to address users who do not hold a committed AWS presence or are just onboarding to the platform. Data is consolidated with Entity Resolution, encodings are stored with Vector Engine, and FMs are fine-tuned over-optimized T5 EC2 instances. Neighboring tasks such as request serving are made possible by Amazon’s other standalone offerings, such as Lambda and API Gateway. To the delight of cloud practitioners growing in experience like us, AWS has the AI ambitions of any user covered.

• Looking wider, we are happy to see AWS actively keep up with the AI frontier. The platform prioritizes staying up to date with groundbreaking models, going as far as conducting its own cutting-edge research to back its in-house model offerings. Also, though AWS platforms are commonly associated with managed software services, AWS is not letting up on progressive hardware improvements. This reinforces the complete cloud stack, empowering even further development possibilities.

• Besides technical excellence, Capital One’s TIP empowers the development of business-oriented skills, such as presentation. Since our first demos were only weeks in, we have critically considered how to address a key challenge: keeping presentations both insightful and engaging. These notions resurfaced when hearing the announcement of Quicksight Q. We were excited to hear how AWS Quicksight compliments the human analyst in the loop, utilizing AI to craft not just presentation materials, such as diagrams and slides, but compelling narratives that continue the conversation.

• Finally, we were most exhilarated by the announcement of AWS Bedrock and Agents for Bedrock. Packaging the entire generative AI lifecycle into a managed service enables developers to focus more on end users and less on the busywork, complexity and guesswork necessitated by ground-up AI development. We left the keynote inspired beyond our expectations, babbling with each other on the infinite near-future achievements now made entirely feasible, welcoming and stimulating.

AWS Summit marked a leap forward in the evolution of generative AI and its assimilation within a global ecosystem of cloud software applications. The conference has us hanging on the edge of our seats for more groundbreaking innovations to follow. In all, the soon-to-be initiated members of the AWS toolchain are prime examples of how AWS makes technology accessible to developers and organizations with an at-scale vision. 

In an all-around win, AWS building a better cloud translates to Capital One innovating the modern financial platform, and changing banking for good. Deep down, we are software engineers determined to redefine technological capabilities and their real-world impact. With graduation winding closer and our careers evolving in the years ahead, this one-day event leaves us hungry for more: The conversation of building on the AI Cloud is just booting up.