Weeks 19 - 22
January always sneaks up on me.
Every year, I look up and realize the month is almost over, and I’m not entirely sure where the days went. I’ve been trying to be more intentional with my time this year. Trying is the keyword.
Spring 2026
Spring 2026 is officially underway (I'll be graduating in May, which still feels surreal to say). This semester, I'm taking INFO 6500: Information Science Seminar Series. It’s a one-credit course, but the intent behind it feels expansive. The focus is less on workload and more on intellectual exploration: learning about research outside my immediate interests, practicing critical writing, and actively contributing to an intellectual community.
So far, the speakers have been incredible.
One of the early sessions explored information history through the lens of Thomas Kuhn and the evolution of the field itself. The idea that information history is not synonymous with the Information Age. It’s older and deeply embedded in social, political, and institutional structures. We discussed microhistory as a method, including a case study focused on the United States in 1920. Retelling familiar narratives like agriculture or the Spanish Flu from an information perspective completely reframed how I thought about data, suppression, misinformation, and information-intensive institutions long before modern computing existed.
Another talk that really stayed with me was “Tainted Source Code” by Professor Bryan Choi. It tackled a hard question the tech world is still avoiding: accountability in open-source software. With the majority of modern codebases relying heavily on open source, the discussion around liability, standard of care, and the analogy to tainted food or blood donations was necessary. It made me think deeply about software as a technical artifact and as a professional and ethical responsibility, especially as AI systems become more centralized, more powerful, and more consequential.
Later in the month, we heard from Jason Lucas on trustworthy AI through low-resource multilingual NLP and cybersecurity. The conversation moved beyond model accuracy and into sociotechnical consequences: whose languages are represented, whose voices are amplified, and how safety mechanisms often fail outside high-resource English settings. The discussion on transfer learning, negative transfer, and robustness across dialects highlighted how fragile our systems can be when deployed at scale. Responsible modeling requires historical, cultural, and linguistic awareness.
Grad+ Fundamentals
This semester, I also joined Grad+ Fundamentals, which sits at the intersection of performance development and social-emotional well-being.
The focus is on attention management, motivation, stress patterns, and aligning daily behavior with long-term values. The sessions have been surprisingly practical. It’s made me think differently about productivity as directing attention with intention.
Intention is something I’m actively trying to get better at: doing one task at a time, not multitasking, and actually noticing how much work gets done when I focus on something fully, even for a short period. Human behavior mastery is complex and fascinating, and I’m genuinely excited to dig deeper into it this semester.
Research Recognition: CU Libraries Showcase
One particularly meaningful moment this month was the CU Libraries Collaborative Research Grant showcase, where the tsunami marigram metadata extraction work I contributed to was presented. I’m grateful to have been selected from more than 100 students, and especially thankful to Erik Radio and Aaron Sweeney for trusting me throughout the process.
The work focuses on digitizing and structuring historical tsunami marigram scans, transforming fragile archival material into usable, reproducible datasets. Knowing that this pipeline will support long-term climate and oceanographic research still means a great deal to me. I’m genuinely grateful to have been part of a process where data engineering directly enables future scientific inquiry. Check out the project here!!
Graduate Course Assistant
This semester, I began working as a Graduate Course Assistant for INFO 5602: Information Visualization. This course meant a lot to me last semester, and getting the chance to return to it from the other side feels special. I’m deeply grateful to Prof. Abel Iyasele for trusting me with this opportunity. Being able to re-engage with the material, support students, and keep that creative side of my work alive has been really nice. I’m loving it more than I expected.
Research & Accessibility: RIO Impact Intern
I’m continuing my role as an Impact Intern with the Research & Innovation Office. Recently, I worked on a website that automatically refreshes every two weeks to surface Pivot-RP funding opportunities, making them accessible to CU Boulder faculty without restrictive access barriers.
Being part of a process that reduces friction and improves discoverability for researchers feels meaningful. I’m especially grateful to Dr. Tanya Ennis, who has helped me sharpen how I communicate, manage time, and stay accountable across moving pieces. Check out the website here!
Boulder Climate Ventures: Amply
Through Boulder Climate Ventures, I joined Amply as part of a climate tech discovery program. The idea of pairing distributed energy storage with AI compute to create localized, revenue-generating infrastructure is ambitious and genuinely fascinating. It sits at the intersection of energy systems, data infrastructure, and incentives, which are areas that are often treated separately but clearly shouldn’t be.
I’m excited to learn more about the energy storage space and explore how AI and data science can play a role here. I’m also really looking forward to working with Walter Richard and Banerjee Pyla and seeing where this collaboration leads.
More broadly, having conversations with people building thoughtfully in climate and applied AI spaces has been amazing. It's great to be around teams that ask hard questions, think long-term, and try to design systems that actually work in the real world.
Parsyl
Outside of grad school, it’s been a little over a month at Parsyl, and I’ve been learning a lot - fast. I’ve started experimenting with n8n automation, building a proof-of-concept that I recently got to demo at an all-company meeting. They’re still very much works in progress, but hearing how much time they could save the team was incredibly rewarding.
I’ve also spent time learning the company’s history, mission, and how data-driven insurance actually works in practice, especially within Claims. I’m grateful to be on a team that values learning, asks good questions, and takes the time to explain the why behind the work. I have a feeling I’m going to surprise myself with how much I grow here.
AI & Complex Insurance Claims
I attended a session on AI and complex insurance claims that helped clarify how automation can meaningfully support high-stakes decision-making without replacing human expertise.
Since I'm new to the insurance space, I thought it'd be a great start to know where the industry currently stands with the usage of AI. It was interesting to see that only ~7% of companies are currently using AI in a way that’s deeply embedded across routine workflows and touchpoints. There’s clearly a lot of room for growth.
The discussion focused on how AI can assist adjusters in navigating severity, ambiguity, and scale, especially in specialty lines where context and judgment still matter deeply. There was an emphasis on balance: using AI to reduce cognitive load and improve consistency, while keeping humans firmly in the loop where nuance and accountability are critical.
Agentic AI & Amazon Bedrock
I attended an AWS workshop on building agentic workflows with Amazon Bedrock, which offered hands-on exposure to how autonomous AI systems are designed and orchestrated in practice. The focus went well beyond chat-based interactions, emphasizing agents that can maintain context, reason through multi-step tasks, and act toward user-defined goals.
Working through the labs highlighted how quickly complexity grows once systems move past single-prompt flows and how important thoughtful decisions around memory, control boundaries, and orchestration become at that stage.
Instead of spending time on setup or resource management, the emphasis stayed on system behavior: how agents plan, how they recover from failure, and how workflows remain predictable, auditable, and safe as autonomy increases.
As part of the workshop, I earned the Amazon Bedrock Workshop badge, which felt like a meaningful checkpoint. It represented building real systems end-to-end, designing agents that plan, reason, and execute responsibly within defined constraints. More than anything, it marked a shift from learning about agentic AI to actively designing with it, helping solidify how these concepts translate from theory into production-ready workflows.
Startups2Students
I also attended Startups2Students, which offered a glimpse into early-stage companies working at the intersection of AI, data, and applied problem-solving. The room is always so energetic.
Companies like BlueVector AI, which applies machine learning to cybersecurity and threat detection, and AcudocX, which works on intelligent document processing, stood out for how directly they translate data science into operational impact.
It was encouraging to see how many paths exist outside traditional big-tech roles, and how much room there is for impact-driven, technically rigorous work in smaller, fast-moving teams. Looking forward to staying connected with these companies!
RISC & Risk Intelligence
I also attended a RISC information session, which provided insight into how the organization approaches risk intelligence through data-driven analysis.
I really liked their emphasis on clarity and responsibility, using data to predict outcomes, to help organizations understand risk in a way that supports better judgment. What I loved most was the people. Everyone I spoke to genuinely loves their work and is motivated by solving problems that help society. There’s a shared belief that no problem is off-limits and no idea is a bad one, and as someone with a lot of ideas, that mindset really resonated with me. I’m looking forward to learning more.
And finally, some personal notes:
Lately, I’ve been reading Life Principles by Ray Dalio, and it’s been nudging me to reflect more honestly on how I make decisions. It’s been less about finding answers and more about asking better questions. Clear goals. Honest diagnosis. Designing around problems instead of wishing them away. Doing what’s necessary, even when it’s uncomfortable.
None of it is new to me. But reading it now made me realize how often I rush past the process and fixate on outcomes instead. I want results quickly. I want things to work immediately. And when they don’t, I tend to skip reflection and move straight to frustration.
I’m trying to slow down enough to actually apply the loop, especially in how I plan my time, approach learning, and respond when things don’t go as expected. Less perfection, more iteration. Less self-judgment, more clarity. Easier said than done, but I’m trying.
On a lighter note, I found this amazing place that serves Mandi, which is basically biryani, but a lot of it. Like… a LOT. It looked incredible and tasted even better. What made it even nicer was meeting some of my undergrad seniors, who are also my seniors here now. It’s surreal to think we’re halfway across the world. Not that long ago, we were just kids trying to get by, and now we’re here, building lives we didn’t even imagine back then.
Boulder was cold last weekend. It snowed a lot. But what I love about this place is how quickly it shifts. It snows one weekend, and the rest of the month is bright and sunny. The snow melts, the trails open up, and suddenly it’s perfect hiking weather again. We somehow get the best of both worlds. Snowy weekends, sunny days, and those sunrises and sunsets that make you stop.
Well, that's been January. February looks packed, hoping to show up, learn, and grow more!
Comments
Post a Comment