Engineering Intelligence: Chuanrui Li and the Convergence of Software and Intelligence

With over a decade of experience in the software industry, Chuanrui Li has closely observed and actively participated in the transformative impact of artificial intelligence on the field of software engineering.

To understand the scale and impact of AI on software engineering, Li believes it is useful to frame the field across three distinct eras: the pre-AI era, the current AI-augmented era, and the future AI-native era.

Pre-AI Era: Ownership of Code and Process

At the beginning of his career, Chuanrui Li worked in an environment where software engineering was primarily centered around writing scalable, maintainable, and clean code. Engineers were responsible for the entire development pipeline—from designing solutions to implementing, testing, and maintaining them. It was a time when engineers functioned like skilled craftsmen, meticulously managing every detail of their systems, akin to supervisors on a factory floor.

During this period, AI was just beginning to attract attention, primarily within the realm of data analytics. Internet companies such as Google recognized that valuable insights could be extracted from user behavior—click patterns, session durations, and ad engagement—while quantitative trading firms focused on deriving meaning from financial data like trading patterns, market sentiment, pricing anomalies, and transaction-level trends.

This wave of interest gave rise to specialized roles like data scientists and machine learning engineers, whose primary focus was to uncover patterns and generate insights from massive, high-dimensional datasets.

The emergence of platforms such as Snowflake and Databricks supported this data-centric paradigm. Snowflake offered a cloud-native data warehouse capable of storing and analyzing vast amounts of structured and semi-structured data with high scalability and performance. Databricks, built on Apache Spark and other open-source technologies, provided a unified platform for big data analytics and AI development, allowing for scalable and collaborative data processing.

These platforms were grounded in open-source ecosystems—Spark, Delta Lake, and MLflow—which empowered modular, flexible, and community-driven innovation. Still, at the heart of software engineering during this period was human intuition, manual coding, and incremental iteration.

Current Era: AI-Augmented Engineering

Today, Chuanrui Li notes that major tech companies have already entered a new phase of engineering—one that embraces AI not just as a tool for data scientists, but as a co-pilot in day-to-day software development.

Google CEO Sundar Pichai revealed that over 25% of new code at Google is now generated by AI, with human engineers reviewing and accepting these AI-generated contributions. Similarly, Microsoft CEO Satya Nadella shared that AI is writing between 20% and 30% of Microsoft's code, depending on the project and programming language.

The rise of large language models (LLMs) like ChatGPT marked a pivotal turning point. Instead of viewing AI as a separate layer of intelligence, engineers began integrating it directly into their development workflows. AI evolved into an augmentation tool, one that didn't replace engineers but enhanced their efficiency and creativity.

Chuanrui Li identifies three core ways AI augments software engineering today:

1. Accelerated Domain Understanding

In highly specialized industries like fintech, where systems are domain-specific and complex, engineers historically spent weeks studying financial concepts or internal documentation before contributing. Now, Li points out, AI tools can rapidly summarize relevant materials and clarify domain-specific requirements, allowing engineers to ramp up much faster.

2. Coding and Workflow Assistance

AI has become deeply embedded in daily engineering tasks. From writing test scripts and addressing code review feedback to implementing modular features, AI significantly reduces the time engineers spend on repetitive, routine tasks. This shift allows developers to focus on higher-order design and architectural decision-making.

3. System Comprehension and Visualization

One of the most daunting challenges for engineers is understanding legacy codebases. In the past, this required weeks of diving into dense documentation and tracing complex dependencies. Now, AI can highlight critical code paths, summarize logic, and even generate visual diagrams that reveal how services and classes interconnect—tools that dramatically accelerate comprehension and onboarding.

With these advancements, Li explains, the role of the engineer is evolving. Engineers are no longer merely executors; they are becoming orchestrators—still responsible for system design and oversight, but supported by intelligent tools that streamline implementation.

Next Era: AI-Native Engineering

Looking ahead, Chuanrui Li envisions a future where software engineering mirrors the evolution of mechanical engineering.

Modern mechanical engineers rarely build machines by hand. Instead, they rely on tools like SolidWorks, AutoCAD, and Ansys to model, simulate, and test mechanical systems in virtual environments. The final designs are passed on to robotic assembly lines and CNC machines that carry out physical production.

Similarly, in an AI-native software world, engineers will focus primarily on high-level system design, logic, and validation. Intelligent models will handle most of the implementation. The software engineer of the future becomes a systems architect, responsible for ensuring that AI interprets and executes their design intent correctly and reliably.

In Fintech: Engineers as Model Designers and Integrators

This paradigm shift will also collapse the traditional barrier between quantitative researchers and software engineers. Engineers will take full ownership of both model design and system integration.

Li likens this transformation to circuit design. Engineers will need a deep understanding of how each AI model functions and how to interconnect them to build cohesive, intelligent systems. Instead of relying on manual statistical modeling, engineers will guide AI systems to generate, test, and optimize financial models based on defined objectives.

Quantitative firms will no longer require large teams of statisticians manually crafting formulas. AI will handle much of that work—rapidly generating models with superior adaptability and performance—while engineers oversee and refine the process.

Chuanrui Li concludes that software engineering is undergoing a profound evolution—from manual, hands-on coding to intelligent orchestration.

The modern engineer will no longer just write code. They will design, validate, and manage intelligent ecosystems—systems that can adapt, learn, and improve themselves. In this new era, success will belong to engineers who blend deep technical knowledge with creativity and a strong sense of purpose.

The future, as Li sees it, belongs to those who don't just use AI—but understand when, how, and why to use it.