Day 2 Online Workshops

AI is evolving beyond simple chat experiences into intelligent systems that collaborate with humans. This session explores how to design the next generation of AI interfaces by combining voice interaction, visual context, dynamic widgets, and agent-driven experiences. Learn how to build AI experiences that feel natural, understandable, and aligned with how humans think, work, and interact with technology.

For decades, QA defined itself by certainty—pass or fail, green or red. But that line is shifting: AI has absorbed the rote work, roles now blur between tester, engineer, and coder, and our tools increasingly estimate rather than declare. For a profession built on certainty, this can feel like losing sight of land.

This talk reframes that anxiety as a horizon—not a destination to reach, but a line to steer by—and argues that QA’s most durable skill was never the green checkmark, but adaptability. We’ll explore how uncertainty can be reframed as flexibility, and flexibility as navigation—so you leave not with a fixed map, but with a way to keep moving toward what’s next, even when the ground won’t hold still.

As organizations accelerate their adoption of AI, analytics, and modern data platforms, one challenge remains constant: transforming raw, fragmented, and inconsistent data into trusted, analytics-ready information. Data engineers and business users often spend a significant portion of their time cleaning, mapping, and preparing data before it can generate meaningful insights.

This session introduces Fabric Data Transformer (FDT), a modern data transformation solution designed to streamline data preparation within the Microsoft Fabric ecosystem. FDT empowers organizations to simplify complex data transformation processes through intuitive, scalable, and efficient workflows, enabling faster delivery of analytics, reporting, and AI initiatives.

Participants will gain insights into the common challenges of data transformation, the role of Microsoft Fabric in modern data architectures, and how FDT helps bridge the gap between raw data and business value. The session will also showcase practical use cases, best practices, and real-world scenarios demonstrating how organizations can accelerate their data journey while reducing complexity, cost, and time-to-insight.

Whether you are a business leader, data professional, IT manager, or technology enthusiast, this talk will provide a practical understanding of how modern data transformation platforms can serve as a foundation for successful AI and data-driven decision-making.

By the end of the session, participants will be able to:
Understand the challenges of modern data preparation and transformation.
Appreciate the role of Microsoft Fabric in a unified data platform strategy.
Identify opportunities to improve data quality and analytics readiness.
Learn how Fabric Data Transformer simplifies data transformation workflows.
Explore practical applications of transformed data for reporting, business intelligence, and AI initiatives.

In the evolving landscape of interactive media, the line between simulation and reality is blurring. This talk explores the pioneering R&D project to be funded by DOST-PCIEERD, titled “GameDev x Computer Vision,” where we attempt to apply game development techniques in the field of computer vision and AI.

Beyond the project details, this session will share how our initiatives secured a government‑funded grant, showcasing a successful blueprint for collaboration between academia and industry—including both traditional software firms and game development studios—to push the boundaries of what’s possible in the local tech ecosystem.

We’ve all seen the flashy demos, but the reality of using AI agents in enterprise codebases is setting in: architectural drift, silent regressions, and a mountain of ‘AI spaghetti code.’ The industry is hitting collective AI coding fatigue. In this talk, we step away from the hype and look at the discipline.

In this session, discover Spec-Driven Development (SDD)—a pragmatic methodology that shifts your role from writing raw syntax to engineering structured intent, using machine-readable specifications to force AI agents to respect your architectural guardrails.

Modern web scraping requires navigating strict bot defense systems like Akamai and reCAPTCHA. Extracting data reliably at scale demands an architecture that accurately mimics human behavior for automated problem resolution.

This session outlines the technical requirements for building resilient, high throughput scraping pipelines capable of operating undetected overtime.

Technical Breakdown
– Architect Execution Pipelines: Evaluate Headless versus Headed Chrome with Xvfb, direct API endpoint requests, and extension-based scrapers. Implement stealth automation using frameworks like Playwright, Patchright, and curl_cffi.

– Browser Searching: Use locally hosted services like duckduckgo search and SearXNG to save on Google Search or Serper credits

– Bypass Automated Verification: Deploy audio processing pipelines using ffmpeg, Faster Whisper, and Google STT alongside visual models like CLIP to programmatically solve verification challenges.

– Extract High-Value Data: Leverage Optical Character Recognition (OCR) engines like PyMuPDF, Tesseract, and RapidOCR strictly for data extraction from documents and images.

– Manage Digital Identities: Configure residential proxy rotation, spoof granular browser fingerprints to execute session trust-building strategies.

– Orchestrate and Scale Behavior: Transition your architecture from a single worker to multi-threaded concurrency, and ultimately scale across multi-VM deployments.

🏆 The Challenge

Data scraping often operates in a legally grey area. Explore this space practically by putting your new skills to the test on a special practice target website designed for you to tackle in your free time at the end of the session. Compete to see who can navigate the architecture, bypass the defenses, and retrieve the most complete information in the fastest time.

AI agents are everywhere, but the real disruption isn’t the agent itself—it’s what they’re doing to us. As AI takes over repetitive coding tasks, a new role is emerging that can’t be automated away: the product engineer. This is not the death of engineering; it’s an upgrade.

Product engineers still code, but they think in terms of outcomes, users, and business value. The ones who thrive won’t be those who wrote the most functions, but those who understood the why behind every feature they shipped.

This session explores how AI is reshaping the identity of engineers, why product‑driven thinking is becoming essential, and how developers can evolve from function writers to outcome shapers in the age of intelligent automation.

The general trend in LLMs is toward ever larger systems, typically designed for hyperscalers. This talk takes the opposite approach—running local agentic AI on low‑end machines by tailoring solutions to the right tasks.

While engineers have automated much of the world, lawyers and corporate teams in the Philippines remain buried under flat PDFs and fragmented tracking systems. This session brings the perspective of “The Lawyer Tired of Manual Work,” turning everyday frustrations into actionable requirements for developers.

We’ll break down the “Holy Trinity” of contract pain points—Review, Tracking, and Management—and show how LegalTech can move beyond generic AI applications to solutions that truly matter. Unlike the classic startup mistake of building first and hoping for adoption, in LegalTech the demand is clear: when developers build what domain experts ask for, the market follows.

With over 75,000 active lawyers and 578,000 SEC‑registered corporations in the Philippines alone, the opportunity is massive. This talk challenges developers to push past surface‑level automation and create tools that transform legal practice into something faster, smarter, and genuinely impactful.

What does it take to build systems that continue to function when connectivity is unavailable?

Modern applications often assume reliable internet access and cloud infrastructure. This talk follows the speaker’s ongoing exploration of local-first system architectures for disconnected and resource-constrained environments. This session will share experiments, design decisions, and lessons from exploring offline‑first, edge processing, and hardware‑aware systems.