# Somya Anand

> I design and build ML systems for better thinking.

## Articles

- [The World is the Test](./writing/2026-03-02-the-world-is-the-test/index.html.md): Bruegel's Netherlandish Proverbs — a world of agents, each acting on their own logic
- [The Era of Personalized Software](./writing/2026-01-22-the-era-of-personalized-software/index.html.md): Illustration of builders creating personalized tools
- [From Reading Rut to Reading System](./writing/2026-01-15-reading-rut-to-reading-system/index.html.md): Last year, stuck in a reading rut, I picked up Michiko Aoyama's What You Are Looking For Is in the Library. It's a quiet book. Five people in Tokyo, each struggling with something, each finding their way to a community library where a librarian named Sayuri Komachi hands them exactly the book they didn't know they needed.
- [Finding My Writing Voice](./garden/writing-voice/index.html.md): Prompts and tools I use to develop my writing voice and make my creative ecosystem better.
- [](./writing/2026-01-11-why-i-read-fiction/index.html.md): I found my way back to reading last year through fiction.
- [](./writing/2026-01-10-co-evolving-ai-systems/index.html.md): In some of my recent reading, I found myself returning to an idea from Thinking, Fast and Slow: that intelligence isn’t just about reasoning well, but about bringing the right memory or context to mind at the right time. Much of what we call judgment comes down to what gets surfaced — and what stays buried — in a given moment.
- [Evaluating Work in the Age of AI: A Manager's Perspective](./writing/2025-12-10-evaluating-work-age-of-ai/index.html.md): Why clarity matters more than ever, and how to bring AI-shaped work into performance conversations.
- [Using an LLM as a Rubber Duck for Ideas](./writing/2025-11-18-llm-rubber-duck/index.html.md): One of their most underrated uses is in the earliest stage of engineering projects.
- [I Stepped Away](./writing/2025-10-31-i-stepped-away/index.html.md): Growth doesn't always mean moving forward — sometimes it means standing still long enough to listen.
- [Rewiring My AI Mindset: Hooks, Agents, and Hands-On Learning](./writing/2025-07-26-rewiring-ai-mindset/index.html.md): How AI Engineering by Chip Huyen helped me connect the dots and build better.
- [My Experience Building a People-First Team](./writing/2025-07-02-building-people-first-team/index.html.md): Every mid-year calibration brings me back to the same question: Am I building the kind of team I'd want to be part of? Three years into this role, the answer is clearer than when I began, but it's still evolving. I've made mistakes, learned from my team, and continually tried to build a team that is not only high-performing, but also deeply human.
- [Lifecycle of a Bug](./writing/2021-06-05-lifecycle-of-a-bug/index.html.md): A few days back, I was working on crucial client data, and my ML model was performing as expected, but the final results were really off. I checked cross validation scores to see if this is a case of overfitting, I ran my model on a random sample just to make sure that the data distribution remains the same, I went through all model evaluation parameters, checked all the feature correlation, went through rounds of hyper parameter tuning, error analysis; I checked everything I knew, but I could not find the problem.
- [Design Data to Text Generation Hybrid Templating System Using SimpleNLG](./writing/2020-12-08-data-to-text-simplenlg/index.html.md): About more than a year ago, I wrote about designing a hybrid templating engine using natural language generation library (SimpleNLG). I was really surprised a lot of people found the article helpful and asked for a hands-on tutorial to use SimpleNLG for data to text conversion. I have finally put together a simple tutorial to generate sentences using SimpleNLG.
- [Why Are Code Reviews So Important?](./writing/2020-08-04-code-reviews/index.html.md): As I work as an R&D Engineer, I am getting accustomed to juggling my way between data science and software engineering. I can safely say that I have learnt a lot of best practices through effective code reviews. Whether it is your first code review or the 1000th, each code review helps us to grow a step further.
- [What Did I Learn in 2019?](./writing/2020-01-13-what-i-learned-2019/index.html.md): If I have to pick up the best month, January is probably my personal favorite. A new fresh year but a beautiful nostalgic past. Some good moments and some bad. As I have grown one more year older and one more year in the industry, I wanted to go back and see why was 2019 a very big year in my life?
- [Smart Fare calendar system](./writing/2017-09-15-fare-calendar-system/index.html.md): Design a simple fare calendar system in order to cluster different price range in respective air route to get a better understanding to price variation throughout the year. Instead of having just minimum, maximum and median airfare to represent three classes of fare distribution in fare calendar, we can design a smart airfare categorization using clustering algorithms.
- [Fare Trends](./writing/2017-09-01-fare-trends/index.html.md): Pricing for airlines is one of the most volatile and dynamic metric in the travel industry. No wonder, it can drive potential customer really anxious before buying an airline ticket. Thus there is a great need for proper fare analytics and fare trend visualisations for better user experience and hassle free booking.
- [Role of emotions in online communities: An agent based approach](./writing/2017-08-11-agent-based-modelling/index.html.md): We tend to analyze how people express their emotion in online communities using an agent based approach. We developed an agent based model to see the impact of emotion in integration and disintegration. Further we check the effect of anonymity in the behavior of agents in online communities.
- [Supervised approach for nuclei segmentation](./writing/2017-08-11-supervised-nuclei-segmentation/index.html.md): Segmentation is an important stage in almost every problem involving digital image analysis. In image segmentation we aim to partition the spatial domain of the image, thereby delimiting the region of interest which correspond to the target objects in the concerned image analysis.
- [y2038 issue in linux kernel](./writing/2017-08-01-y2038-problem/index.html.md): The Year 2038 problem or commonly called as y2038 issue is an issue for computing and data storage situations in which time values are stored or calculated as a signed 32-bit integer, and this number is interpreted as the number of seconds since 00:00 UTC on 1 January 1970 (the epoch).[1] Such implementations cannot encode times after 03:14 UTC on 19 January 2038.
- [Fare alert system](./writing/2017-05-12-fare-alert-system/index.html.md): Purchasing a flight ticket, though a simple and quick affair, causes lot of anxiety to many travellers even today due to the fact that airlines frequently change fares. Due to this dynamic nature of airfare fluctuation, fare alert system plays a key role in customer engagement and better user experience in travel industry.
- [User classification](./writing/2017-03-11-user-classification/index.html.md): Based on the activities performed by the user in travel marketplace, we aim to classify them with certain probability that how likely is the user to book a ticket.
- [Coccinelle](./writing/2016-08-11-coccinelle-usages/index.html.md): Coccinelle is a program matching tool which allows programmers to easily write some complex style-preserving source-to-source transformations on C source code, for instance, code refactoring. There are useful online resources which document this tool very well, however it is slightly difficult for a beginner to understand these concepts through these materials. So this is my attempt to explain the details of this tool in more convenient way.

## Other Pages

- [Q1 2026 Reading Syllabus](./garden/reading-syllabus/index.html.md): A structured reading plan across four domains: classic literature, deep fiction, neuroscience, and applied engineering.
- [Somya Anand](./index.html.md): Building AI systems inspired by human thinking—memory, dialogue, and learning through use. Curious about NLP, personalization, lifelong agents, and deeply human tools. This is my digital garden. If you’re exploring interesting problems in ML or AI, I’d love to hear about them.
- [Writing](./writing/index.html.md): Thoughts on AI, engineering, management, and the spaces where they intersect.
- [About Somya](./about/index.html.md): Applied scientist, builder & writer. I tend to my digital garden here.
- [Garden](./garden/index.html.md): A digital garden is a place for ideas to grow at their own pace. Some are seedlings—just planted, barely formed. Others have been here longer, branching in unexpected directions.
- [Reading](./reading/index.html.md): I read widely and slowly. Fiction for joy, non-fiction for understanding, and technical books for craft.
- [Reading Companion](./garden/reading-companion/index.html.md): An MCP server that helps you set reading goals, discover books, and deepen learning through reflection.
- [Learning in Public](./garden/learning-in-public/index.html.md): Learning in public means sharing your process, not just your conclusions. It means being okay with being wrong, because being corrected is how you learn faster.
- [Systems Thinking Notes](./garden/systems-thinking/index.html.md): Fragments from reading about complex adaptive systems, organizations, and how things change.
- [Book Log](./reading/log/index.html.md): Everything I've read, in rough chronological order. A personal record, not a recommendation list—though I'm happy to talk about any of these.
- [Currently Reading](./reading/now/index.html.md): A multi-generational story set in Kerala, spanning from the early 1900s to the present. It's long and dense and beautiful. The kind of book where you need to occasionally stop and just sit with what you've read.
- [On AI and Feedback Loops](./garden/ai-feedback-loops/index.html.md): Quick thoughts on designing systems that learn from use.