Category: Management
-
The Rise of Small Language Models and Open-Source AI: Opportunities for Asia-Pacific Data Scientists
Beyond the Billion-Parameter Arms Race For years, the narrative in artificial intelligence has been dominated by scale—more parameters, larger datasets, and ever-increasing computational budgets. This race towards massive, centralized models created a significant gap. For most organizations, especially in the dynamic and diverse Asia-Pacific (APAC) region, the costs, infrastructure demands, and privacy implications of deploying…
-
Data Governance in the Age of AI Agents: Lessons from Asia’s Regulatory Landscape
When Your AI Assistant Becomes a Compliance Liability The rise of agentic AI systems—autonomous workflows that can execute tasks, make decisions, and interact with other software—marks a watershed moment for data governance. As I’ve written before regarding professional integrity in data science, control over data is the cornerstone of trust. However, an AI agent that can…
-
Building AI Factories: How Hong Kong Companies Can Create Scalable GenAI Infrastructure on a Budget
From Cost Centre to Competitive Engine: The “AI Factory” Mindset For Hong Kong’s dynamic businesses, the promise of Generative AI (GenAI) is tempered by a harsh reality: the perceived high cost and complexity of building a robust, scalable infrastructure. Many companies find themselves trapped in “Pilot Purgatory,” where impressive prototypes—like the autonomous agent systems we’ve…
-
Agentic AI in 2026: From Hype to Real-World Deployment in Asian Enterprises
Introduction: Emerging from “Pilot Purgatory” Just two years ago, in my 2024 post on “Pilot Purgatory in Machine Learning,” I discussed the frustrating gap between promising prototypes and deployed production systems. Today, as we examine the state of Agentic AI in 2026, we witness a remarkable transformation. The landscape has evolved from isolated experiments with tools…
-
Pilot Purgatory in Machine Learning: Why Most Models Excel in Prototyping but Fail to Deploy in Production
The POC-to-Production Gap: A Persistent Challenge in Applied ML As data scientists, we’ve all encountered the frustrating reality of “pilot purgatory”—where promising proof-of-concept (POC) models deliver impressive offline performance but never make it to production. Industry reports highlight the scale of this issue: estimates suggest that 70-80% of ML projects fail to reach production deployment,…
-
The Misconceptions of LLM: Is a Large Model Really Omnipotent?
In recent years, with the rapid development of large language models (LLMs), many corporate executives have been eagerly embracing this technology, believing it to be a panacea for all problems. Since early 2025, the rise of DeepSeek has further fueled market enthusiasm, especially in Hong Kong, where many enterprises have begun massive investments in LLM-related…
-
Changes in Moral Standards and Current Workplace Challenges
In recent years, many business leaders and management teams have observed a sharp decline in work ethics, sense of responsibility, and dedication. Particularly after COVID-19, project execution efficiency in the Asian market has been significantly affected, mainly due to employees’ uncooperative attitudes and indifference toward project progress. Internally, companies are also facing challenges with younger…
-
From Blocker to Builder: Transforming IT to Fuel Business Innovation
Introduction In today’s fast-paced business environment, technology is the backbone of growth, innovation, and efficiency. Yet, many organizations find their IT departments caught in a reactive mode, focused primarily on maintenance, security, and existing infrastructure, leaving little room for innovation or collaboration with other business units. As a result, IT can be viewed as a…
-
Standing Firm on Professional Integrity: The Crucial Role of Data Scientists in Data Warehousing and Governance
In the realm of data science, maintaining professional integrity is not just a matter of ethics; it’s a cornerstone of effective and meaningful work. Recently, I encountered a situation during a data warehouse project that highlighted the importance of this principle. I authored a “Data Gap Analysis Report,” where I identified several internal data errors…
-
Challenges and Strategies for Improving Efficiency in Data Science Project Teams
In today’s data-driven world, data science projects are gaining increasing attention. However, despite the high technical expertise of many team members, inefficiency often plagues project teams, turning them into a disorganized group. This inefficiency not only delays project timelines but also potentially impacts the final outcomes. This article will discuss common issues within data science…