Company News

In my experience with recruitment, the job market for Business Analytics (BA) majors in North America is a mix of highs and lows. Not long ago, I helped a mid-sized e-commerce company find a business analyst. We received 237 resumes, and the candidate we selected didn’t have an outstanding academic background, but what stood out was his real-world experience working on an industry-level practical project, which impressed us. When it comes to salary, the market can feel a bit unpredictable. I interviewed a fresh graduate from UBC last month who was asking for a $90K base salary, but we ended up offering $65K, which was already a strong offer. However, it’s interesting that those with 1-2 years of experience often end up with 40%-50% more in salary than fresh grads. I know a young woman who started at $58K in Vancouver and, after switching to Toronto, saw her salary jump to $83K. The actual job responsibilities may not always align with what students expect. The senior-most analyst on our team recently spent three days just cleaning data—sorting through messy sales numbers. In reality, modeling and analysis only take up about 20% of the job, while most of the time is...

Job hunting in the U.S. is like running a marathon—if you start late, you’ll fall behind. For international students graduating in 2026, the fall recruiting season of 2025 is a crucial period that could set the stage for their careers. Missing this window could mean putting in double the effort to catch up. Every January, while most students are still enjoying their winter break, the smart ones are already getting ready. A resume isn’t something you just write; it’s something you perfect. Ask a few industry professionals to review it. Their feedback can often open your eyes. For instance, a friend studying finance was able to turn a small, unnoticed club activity into "leading a 10-person team to organize an inter-campus event," which eventually helped her secure an interview at JPMorgan Chase. March and April are the peak months for internship applications. Last year, a computer science student with a GPA of only 3.4 made it onto Amazon’s internship list by showcasing three complete projects on GitHub. While landing an internship with a big company is great, hands-on experience at smaller companies can also make you stand out during the fall recruitment. During this time, you should also start networking....

Getting a Google referral is indeed a rare opportunity, but how far it can take you still depends on your own abilities. Having spent several years in Silicon Valley, I’ve seen many people successfully land a job at Google through referrals, but I've also seen just as many get rejected even with a referral. At the end of the day, a referral is just your ticket in—not a guarantee of success. Google’s resume pool is vast and deep. When I referred a friend last year, the HR team privately told me that they receive about 300 resumes per day for an entry-level position. A regular application is like throwing a stone into the ocean, but a referral at least makes your resume float on the surface. However, don’t think that guarantees success—one interviewee I know said that out of the last 10 referrals they received, only one made it through. Cultural fit is more important than most people realize. Google has an unspoken standard in their interviews: not only do you need to be smart, but you also need to be "Googley." Last year, a technical expert was rejected because he was too focused on working alone. The interviewer’s exact...

Machine learning is becoming the new darling of the U.S. job market, and graduates with a Master's in this field are in high demand. A quick glance at job boards will show that from Silicon Valley tech companies to Wall Street investment banks, to research departments in major hospitals, everyone is scrambling to hire machine learning talent. Tech companies are the obvious employers. The apps and websites we use every day all rely heavily on machine learning. Recommendation algorithms predict what products you might like on e-commerce sites, while speech recognition helps your phone understand what you’re saying. All of these technologies require skilled professionals to develop and optimize them. It’s not just big tech firms—now even small businesses, like your neighborhood bakery, are figuring out how to use AI to predict sales for the next day. The financial sector has also transformed. The days when bankers wore suits and ties are long gone; now, the offices are filled with software engineers in T-shirts and jeans. They’re using machine learning models to predict stock prices, detect suspicious transactions, and even automate loan approvals. A friend working at an investment bank told me that now, the time spent discussing algorithms in...

It's 4 a.m. in New York, and the lights are still on in the financial district. I’m sitting at my desk in the data department of an investment bank, going over a quarterly financial report. Suddenly, I notice an odd number fluctuation. The trends look perfect, but something seems off. As I dig deeper, I realize the recent Federal Reserve interest rate hike is completely missing from the analysis. That’s a huge piece of context that the report didn’t account for. I spent the rest of the night reworking the analysis, ultimately helping the client avoid a costly mistake. That moment made me realize: Generative AI is changing things, but it will never replace the judgment and creativity that only a human analyst can bring. Lately, there’s been a surprising trend in hiring: the companies that are most enthusiastic about adopting generative AI are also ramping up their recruitment for data analysts. Just last month, a recruiter from a top Silicon Valley tech company told me that their new "AI Validation Analyst" role had received thousands of applications. This isn’t a fluke. Businesses are slowly recognizing that AI-generated insights are like raw diamonds—they’re valuable but need to be shaped properly....

The AI learning path is like climbing a high mountain, which requires both a clear direction and solid steps at every turn. This path usually begins with Python programming and a foundation in mathematics, which is the starting point for all AI engineers. I remember when I first started, it took me a whole week just to understand the broadcasting mechanism of NumPy and repeatedly debug those matrix operations with mismatched shapes. This is especially true for the foundation of mathematics. When deriving the loss function of logistic regression, that late night when I suddenly figured out the chain rule is still unforgettable to this day. These seemingly dull foundations precisely form the most solid foundation on the AI learning path. The math almost broke me. Linear algebra felt like trying to read a foreign language at first. What finally made it click was applying concepts directly to code. I'd implement matrix operations by hand in Python, then compare my results to NumPy's output. That painstaking process revealed how mathematical abstractions translate to actual computations. Probability was even tougher - I must have drawn hundreds of bell curves before distributions started making intuitive sense. Kaggle taught me brutal but invaluable...

During a hot Pittsburgh summer, I found myself sitting in the underground lab of CMU’s Gates Computer Center, staring at the blinking code on my screen. My “Introduction to Algorithms” book was dog-eared, and my coffee mug had a thick layer of stains at the bottom. I had no idea that in just three months, I’d be standing under Amazon's iconic glass dome, "The Spheres," in Seattle, signing a return offer. One of the most memorable moments of my application process was the unexpected challenge. I was in the middle of Amazon’s online coding test when the CMU campus network crashed. I grabbed my laptop and ran to the nearest café. With the WiFi barely working, I completed the last two algorithm questions by hand, using pen and paper. My interviewer later told me that it was those scribbled pseudocode notes on napkins, stained with coffee cup rings, that made him recognize my ability to solve problems under pressure. Those napkins, marked with coffee stains, became the most real part of my application. On the first day of the internship, I was hit with the work culture shock. My badge wasn’t even activated yet when I was pulled into an...

In North American AI engineer interviews, core questions that frequently appear usually revolve around several key dimensions. Mastering the response skills to these questions can significantly increase the pass rate. Algorithm implementation questions are almost always included. Questions such as writing the backpropagation algorithm by hand or implementing regularized logistic regression directly test mathematical coding skills. When solving these problems, one should explain the changes in matrix dimensions and the gradient calculation process while writing the code. The mathematical derivation of the bias-variance trade-off is a frequently asked theoretical question in machine learning. The best answer should combine specific algorithms, such as how the tree depth in random forests affects the two, and extend to the analysis of the learning curve during actual parameter tuning. Deep learning architecture design questions such as "How to optimize the memory usage of Transformer on long texts" have almost become standard, requiring a full-stack optimization thinking from sparse multi-head attention to gradient checkpointing techniques. Frequent engineering practice questions include the design of feature storage solutions. When answering, one should compare the impact of the update frequency of real-time features and offline features on model performance and provide scenario-based solutions. Recommendation system problems in system...

Thinking of transitioning to AI but don’t know where to start? Don’t worry. In North America, the AI industry values your ability to learn and gain hands-on experience more than your starting point. My own experience of transitioning from mechanical engineering to AI taught me that with the right approach, you can go from zero knowledge to job-ready in just one year. What tech companies in North America care about most is your ability to solve real-world problems. Instead of getting stuck on complex theories, it's better to start by building a working project. I recommend starting with Python, but don’t just stop at learning the syntax. I know someone who, in his first week, used Pandas to analyze his credit card spending data, and by the second week, he built a spending prediction model. This project-based learning approach is far more efficient than taking a structured course. Mathematics is indeed the foundation of AI, but you don’t need to become a mathematician. Focus on mastering key concepts like matrix operations in linear algebra, conditional probability in probability theory, and the gradient concept in calculus. Here’s a really effective method: for every new math concept you learn, immediately implement it...

I’ve seen a lot of people take the wrong approach to learning data structures and algorithms (DSA). One student from last year stands out. He solved over 500 LeetCode problems but still got stuck when a variant of a problem came up in an interview. The issue was that he focused on solving a high number of problems without truly understanding the underlying concepts. To me, learning DSA is like building a house — if the foundation is weak, no amount of problem-solving will help. I suggest starting with the basics: arrays and linked lists. Make sure you understand their core properties. For example, random access in an array is O(1), but insertion and deletion are O(n), whereas linked lists are the opposite. These characteristics determine how you use them in practice. One of my students had to frequently insert data in the middle of a structure for a project. At first, he used arrays, which performed poorly. After switching to linked lists, his program sped up significantly. This kind of hands-on experience is much more valuable than memorizing definitions. Dynamic programming (DP) is a challenge for many. One student tackled it by first looking at the Fibonacci sequence’s recursive...

Algorithm interviews are a major part of the hiring process for many tech roles. While many candidates feel overwhelmed, the key to success is preparation. Here’s a guide to common algorithm questions and tips on how to prepare effectively. Array and String Problems Array and string problems are common in interviews. A classic example is the "rotated array" problem, which checks your ability to manipulate arrays. "Two sum" is another simple yet important question that tests fundamental coding skills. For substring-related problems, such as finding the longest substring without repeating characters, the sliding window technique is a great approach. Linked List Problems Linked list problems are frequently part of interviews. "Reversing a linked list" is a basic but important problem that forms the basis of many more complex questions. Cycle detection in a linked list is another common topic, and the fast-slow pointer technique is often applied here. "Merging sorted linked lists" tests both your understanding of linked lists and your ability to come up with efficient solutions. Binary Tree Problems Binary trees are another popular topic in interviews. Be sure to know the different tree traversal techniques: preorder, inorder, and postorder, both recursively and iteratively. Questions like "finding the...

In a cafe in the San Francisco Bay Area, I heard a conversation between two engineers with my own ears: "Yesterday, after attending Google's system design round and being asked how to design a distributed file storage system, I suddenly got stuck..." Such scenes are not uncommon in the North American tech circle. As someone who has been through it all, I am well aware that the ability of system design is indeed a key threshold for promotion to senior engineer. I remember when I participated in the system design interview for the first time, my palms were all sweaty. The interviewer asked me to design a real-time chat system. I piled up all the components I knew at once: WebSocket, Redis, Kafka... As a result, I was left speechless when asked. Later, when I worked as a technical interviewer at Facebook, I came to understand where the problem lay - what we need is not the accumulation of technical terms, but the systematic thinking for solving problems. My Indian colleague Ravi has an excellent learning method: Every Saturday morning, he would pull me to practice the system design in front of the whiteboard. We will randomly select a common...