Chatiya Shah

The very first step for me was not jumping into applications, but asking the right questions to myself. Self-talk was crucial. I had to be brutally honest about my goals, readiness, and expectations before moving forward. Here are the questions I asked myself, grouped into the areas that mattered most: Personal Readiness Why do I even want to pursue a master's degree? Do I truly want to commit to an additional two years of study? Am I financially prepared to support both my education and living expenses during this time? Am I ready to live away from family and friends, and adapt to a completely new culture if needed? Am I mentally prepared to face rejections or failures during the application and job search process without losing focus? What kind of life do I want, a stable, comfortable career path, or a more ambitious, high-risk, high-reward journey? How will this master's degree fit into my long-term career goals? How do I plan to compete with candidates who already have work experience? What unique qualities or experiences do I bring that differentiate me from others? Academic Choices Should I go for a broader Computer Science degree to keep my options open, or choose a specialized field aligned with my passion? What specific skills or knowledge gaps am I hoping to fill through my master's degree? What kind of learning environment suits me best, research-intensive, industry-oriented, collaborative, competitive, or a mix? Which universities align with my goals, and why? How significant is the people around me (diversity, ambition, mindset) when choosing a university? Now, there’s no single "right" answer to all these questions, but it’s important to prioritize what you want. It’s okay if you don’t have all the answers at the beginning. What matters is recognizing that these questions exist and being honest about where you stand. For example, when I first asked myself what unique qualities set me apart, I didn’t have a clear answer either. But simply knowing that this was an important gap helped me focus on where I needed to grow. When I started, I didn’t have answers to most of these questions. I just made the decision to begin. Slowly, through experiences, exploration, and a lot of self-learning, things started becoming clearer. My first priority was simple: become really good at what I do. So I went back to basics, I focused on learning how to code, and gradually built a strong foundation in Machine Learning and Data Science.

Once I had a clearer understanding of what I wanted, it was time to act on it. This phase wasn’t about ticking off items on a checklist, it was about building a story that genuinely reflected who I was and what I was aiming for. My journey began with the Machine Learning Club at college. It helped me connect with seniors, learn the fundamentals of machine learning through weekly tasks, and eventually co-author a research paper with them and a professor. As students, we had little idea of how to approach academic writing, but I was lucky to have seniors who guided me throughout. Eventually, I took on the role of Machine Learning Head and had the opportunity to return the favor by mentoring others. Since I was sure I wanted to pursue a career in Machine Learning and Data Science, I focused on getting practical experience. I interned as an Artificial Intelligence Intern at Infiheal, where I worked on a production-ready mental health chatbot using data sourced from psychologists. This was my first hands-on experience with large language models and deploying systems on cloud platforms. After that, I joined IIT Patna as a Research Intern, working part-time on fraud detection in Ethereum transactions using graph neural networks. During the summer after my sixth semester, I interned at Avrio Energy as a Data Science Intern. There, I worked on data pipelining and scaling while contributing to energy efficiency solutions across 70+ McDonald’s outlets in 22 countries. The most impactful phase of my journey, both in terms of data science and personal growth, came through my work at SimPPL. I started as a Research Intern, working on assessing political bias on Truth Social, the platform launched by Donald Trump after being banned from Twitter. At that point, I knew I wanted to dive deeper into data science, but I struggled to find the right opportunities and like-minded collaborators. Being from a CS background, I didn’t have many peers focused solely on core data science. SimPPL gave me that space, an environment where everyone was passionate about the same domain. I was the youngest member of the team, and I was grateful to learn from people more experienced than me. One thing I realized through this journey is that while many people know how to write code for ML models, very few know how to ask the right questions or formulate the right problem statement. That’s what truly sets impactful work apart and it’s something our founder, Swapneel Mehta constantly encouraged me to focus on. While working on the paper, I collaborated with international researchers, including a PhD student from CMU and Swapneel, who is a PostDoc at MIT. Honestly, even if I had gone the corporate route, I doubt I’d have gotten the chance to work with such a team. We eventually published our paper at an A grade conference, something I never imagined achieving during undergrad. Another huge opportunity that Swapneel gave me was to Lead the development of the MVP for an AI-powered Analytics Platform, scaling it to pilot with grants from Mozilla and Google (ExploreCSR). Our goals were to understand loyal audience behavior using analytics data, monitor author performance across categories to expand reach and extract long-term value from decades of user interaction data. We successfully delivered a pilot to New York Public Radio and expanded to LION Network members. This project also gave me a new perspective beyond tech. I worked closely with Dhara Mungra, our program manager, who completely changed how I approached problems. My instinct was always to use complex technical solutions, but Dhara taught me that simplicity often solves better. I never saw myself as someone who could confidently speak for a product, but SimPPL changed that. We regularly met with potential clients, journalists, and industry experts who helped us assess the product’s viability. While many would worry about putting their reputation on the line, Dhara and Swapneel made sure to make me step up, learn how to present myself and lead meetings, and that made all the difference. Working on the product gave me two key takeaways: first, it gave me something meaningful to talk about in conversations and interviews; second, it opened doors to network with industry experts I wouldn’t have otherwise met.

Once I had clarity on my direction and experiences to speak about, I shifted focus to the application components. While basic formats for these are easily available online, here’s what I personally found most important beyond the templates: Resume: Story Over Checklist When working on the resume, it's easy to get caught up in showcasing past experiences, crafting impactful bullet points, and perfecting the formatting. But in the process, we often lose sight of the main objective: to clearly communicate what we specialize in. Instead of simply listing skills and projects, focus on telling a coherent story. Remember, having an internship at a well known company is valuable, but what matters more is the actual work you did and the impact you created.

School Selection: Finding the Right Fit One major piece of advice I’d offer is to start early, ideally 5–6 months before the application deadlines. Begin by identifying programs that align with your interests and start tailoring your profile accordingly. Shortlisting universities is often confusing, and ranking alone isn't always the best metric. Here’s what I did instead: I created an Excel sheet and added columns like college, degree, requirements, cost, coursework, location, job prospects, TA/RA opportunities. There’s no single “best” university, it’s about finding the one that aligns with your profile and career goals. Another key step: talk to current students. I reached out to people already in the programs I was interested in to understand the actual pros and cons. The perspective of someone living that experience is far more valuable than outside assumptions.

SOP Strategy: Connecting Curiosity to Career Remember the self-reflection questions I mentioned earlier? The answers to many of those are essential when writing your Statement of Purpose. Tailor each SOP for the specific university to avoid generic writing. Show who you are, what you’ve done, and what you want to do. Top universities receive thousands of applications, so you need to stand out. Make sure your first paragraph hooks the reader by highlighting your unique motivation. In my case, I connected my interest in behavioral data science to a personal experience. Also, ask trustworthy people, especially those outside your field, to read your SOP. I had my mom review mine. She’s not from a technical background, so if she could understand it, I felt confident the admissions committee would too.

Letters of Recommendation: Choosing the Right Voices If you're applying straight from undergrad, I’d recommend submitting two academic LORs and one professional LOR. Prioritize professors under whom you've done meaningful work ideally, those who've seen you grow over time and can genuinely vouch for your strengths. I took a letter from the professor I worked with on my first research paper through the ML Club. He also taught me Data Science courses and mentored me during the Innovative Product Development project. It’s a plus if the professor’s specialization aligns with your intended field; the admissions committee knows such a person can evaluate you well. For my professional LOR, I chose Swapneel from SimPPL. Having worked with him for almost two years, he had seen me evolve, from a beginner to someone leading a product.

Dealing With Uncertainties: Trusting the Process Many people chase multiple research papers, but I believe quality matters more than quantity. As an undergrad, publishing in a top-tier (A*) conference or Q1 journal is not expected but if you manage it, it gives you an edge. When it comes to standardized tests, no matter how much you prepare, what ultimately counts is your ability to stay calm at the test center. Focus on building that calm.

Lastly, don’t compare your journey to others. The grass always looks greener on the other side. Trust your path, remember how far you’ve come, and believe in the effort you’ve put in.

From the outside, my journey might look well-structured, but the reality was far from linear. Some of the most defining experiences I’ve had came through failures and detours. Looking back, those are the exact moments that shaped both my mindset and trajectory. One of the biggest turning points was being part of DJS Antariksh, my college’s Martian Rover team. Over two and a half years, I participated in six international competitions and contributed to podium finishes in three. But our success was built on a foundation of setbacks. Our first onsite event, the International Rover Challenge 2023 was also my first competition. I worked on the first-ever iteration of the Martian Rover, and watched my seniors run on 1–2 hours of sleep for several days, pouring everything into it. Despite all the effort, we finished 8th. That experience shifted something in me. After that, I realized what it takes to become a master at what you want. I realized that raw technical skill wasn’t enough—what mattered most was how we performed under pressure. That “clutch” mentality of owning the outcome, especially in high-stakes moments, became a cornerstone of how I approached future challenges. The following six months were some of the most intense of my life. I was juggling academics, research, internships, and the rover team, averaging only 4–5 hours of sleep a night for straight six months. My sleep schedule was so messed up that I used to stay awake the entire night and used to sleep while travelling to college and during lectures. I was determined to improve, and made sure no one could ever doubt my technical contributions. But just two weeks before our next competition, the European Rover Challenge 2023 Remote I hit a wall. I was put on bed rest due to burnout. Still, we managed to finish 2nd. That contrast taught me a lesson no classroom ever could: hard work is important, but not at the cost of your health. But more than anything, this phase taught me what it truly feels like to work very hard, something I had never pushed myself to do before. It showed me what I was capable of, and helped me discover my own potential. These experiences led to me being promoted to Coding Head, where I led a 30-member department within a 150-member team. I wasn’t just responsible for getting our rover to work across competition tasks; I also had to mentor and manage. Through this, I started to understand what it means to lead, not just with knowledge, but with clarity, empathy, and accountability. We applied everything we’d learned and secured 3rd place at IRC 2024, my first major leadership competition. Later, we represented India at ERC 2024 in Krakow, competing against MS and PhD students from around the world. We placed 11th overall and were the top-ranked team from India and Asia. You might wonder how being part of a robotics team connects to a career in data science. For me, it made all the difference. It taught me how to stay calm under pressure and deliver results when it mattered most, a skill just as crucial in data science as it is in competitive robotics. More importantly, it helped me understand myself better. This team really opened up my understanding of what tech could look like beyond just projects and hackathons. Representing the country on an international stage, and actually winning, was a perspective-shifting experience and, in my opinion, a strong addition to my profile. Around the end of my second year, I was applying for my first internship through our college fair. I thought I had a solid profile: strong ML fundamentals, rover experience, and a published paper as a second-year student. But I wasn’t shortlisted. The only ML company at the fair had passed on me. Still, I believed I could contribute. So, I created a mini mental health chatbot as a demo tailored specifically for them. On interview day, I simply asked them if they’d be open to giving me a shot, and I got the interview. Competing against peers and even seniors, I landed the internship. They later told me I stood out because I had asked, and because I had built something personal that showed initiative. That moment reinforced a life lesson: not asking is always a no. Another experience that made my journey unique was my time at SimPPL. For someone directly from undergrad, there’s often an unspoken disadvantage compared to those with full-time work experience. But SimPPL helped bridge that gap. I got the rare opportunity to work on a product from scratch, almost like being in a startup environment, while collaborating with international researchers and mentors. It wasn’t just about writing code, I learned how to ask the right questions, how to define the right problems, and how to build things that actually mattered. What made this journey even more meaningful were the people. My seniors, professors, cousins, and internship mentors played a huge role. Swapneel believed in me long before I had any “crazy” experience to show for it. Him and Dhara didn’t just mentor me on technical topics; we had one-on-one conversations about everything from personal websites to how to network better, what to do when moving to a new city, and how to build credibility from scratch. Their mentorship helped me develop not just as a data scientist, but as a person who could speak for a product, lead a team, and carry a vision.

You probably guessed it, I did take a hit on my GPA. If there’s one thing I would definitely change, it’s managing my time better. Even during exam weeks, I often found myself working on projects instead of focusing solely on academics. I prioritized building skills over theoretical knowledge. While I consistently scored well in subjects I was genuinely interested in, my overall pointer was 8.7. Unfortunately, many universities consider GPA a key factor, so that became a hurdle during shortlisting. I’m grateful Columbia evaluated my profile holistically. Another mistake I made was with my GRE prep. Juggling multiple responsibilities meant I could dedicate only eight days of quality study time. While I scored 170 in Quant, my Verbal score was a disappointing 147. I didn’t take the vocabulary component seriously enough, despite being able to comprehend the passages, my limited vocab meant I had to rely on guesswork for several questions. Lastly, while working on SOPs, one thing I overlooked was checking if the university had sufficient funding in my area of interest. Even if you’re highly qualified, if there’s no funding or research focus in your domain, chances of getting in can drop significantly.

Looking ahead, one of the biggest gaps I see in myself is domain knowledge. If I really want to understand human behavior and make better, data-driven decisions, I need to figure out what parameters actually matter, how they change depending on the use case, and how to use them meaningfully. It’s not just about knowing the tools, but about knowing how and when to use them. At Columbia, I want to close this gap by learning from people across fields, joining student clubs, and working with professors who are building things at the intersection of tech and behavioral science. Another big focus for me is figuring out how to use AI systems well in production. Right now, I’m building agents using Agentic AI, but I’m still figuring out how to make them scalable, cost-efficient, and useful in the real world. I want to get better at the engineering side of things, especially for domains like mental health or public systems, where it’s important that the tech actually works at scale. Over time, I want to find the right balance between technical depth and real-world impact. I plan to work in research labs, join interdisciplinary hackathons, and hopefully intern with teams working on AI for social good. But more than anything, I want to keep growing, not just as a data scientist, but as someone who’s grounded, curious, and intentional about the work I do. It’s easy to focus on the admit as the goal. But what really matters is who you become along the way. That’s what Columbia saw, and that’s what I’ll carry forward.