[{"content":"I\u0026rsquo;m Akhil Kota. I study CS and Data Science at Purdue (class of 2027). I spend most of my time on ML research or building software.\nI TA\u0026rsquo;d CS 240. I\u0026rsquo;ve been on the Dean\u0026rsquo;s List, and I\u0026rsquo;ve gotten research recognition through OUR and SURF.\nHow I work For research, I write training code, try models, and check results. On product teams I\u0026rsquo;ve built login, recommendations, scrapers, and dashboards. I like starting simple and improving what works.\nExperience Research Contact ","externalUrl":null,"permalink":"/PersonalWebsite/about/","section":"Akhil Kota","summary":"I’m Akhil Kota. I study CS and Data Science at Purdue (class of 2027). I spend most of my time on ML research or building software.\nI TA’d CS 240. I’ve been on the Dean’s List, and I’ve gotten research recognition through OUR and SURF.\n","title":"About","type":"page"},{"content":"CS and Data Science at Purdue. I do ML research and write software.\nView my work Download resume Currently ML research, Li Group at Purdue Purdue CS and Data Science Building apps and ML tools ","externalUrl":null,"permalink":"/PersonalWebsite/","section":"Akhil Kota","summary":"CS and Data Science at Purdue. I do ML research and write software.\nView my work Download resume Currently ML research, Li Group at Purdue Purdue CS and Data Science Building apps and ML tools ","title":"Akhil Kota","type":"page"},{"content":"","externalUrl":null,"permalink":"/PersonalWebsite/authors/","section":"Authors","summary":"","title":"Authors","type":"authors"},{"content":"","externalUrl":null,"permalink":"/PersonalWebsite/categories/","section":"Categories","summary":"","title":"Categories","type":"categories"},{"content":"Open to research, internships, and product roles. Email is best.\nEmail akhilkota15@gmail.com LinkedIn linkedin.com/in/akhil-kota GitHub github.com/akkota ","externalUrl":null,"permalink":"/PersonalWebsite/contact/","section":"Akhil Kota","summary":"Open to research, internships, and product roles. Email is best.\nEmail akhilkota15@gmail.com LinkedIn linkedin.com/in/akhil-kota GitHub github.com/akkota ","title":"Contact","type":"page"},{"content":"Research labs, startups, and a few industry projects.\nFeb 2026 - Present Machine Learning Researcher Li Group, Prof. Can Li, Purdue ChemE\nWorking on neural methods that use difference-of-convex programming for hard optimization problems. I work on the PyTorch training code, including custom autograd, losses, and how we pick anchors. Goal is a conference paper.\nPyTorchDC programmingCVXPYOptimization Sep 2025 - May 2026 Applied ML Researcher Prabhakar Group, Purdue University\nBuilt tools to segment and track cells in time-lapse microscopy with Cellpose and TrackMate. Fine-tuning cut false-positive microplastic detections by up to 21%. Made the OUR Distinction List and submitted to AAAR.\nCellposeTrackMateOpenCVPython Sep 2025 - Feb 2026 Software Engineer Flow Remote\nAdded Firebase and Google OAuth login and onboarding. Community engagement went up about 35%. Also wrote a small model to predict driver success from location, time, and earnings.\nFirebaseOAuthscikit-learn Jan 2025 - Dec 2025 Founding Engineer Mutually LLC, Bloomington, IN\nBuilt a social events app with React Native and Supabase. Hit 1000\u0026#43; beta users. Wrote a recommender that raised weekly active usage by 40%.\nReact NativeSupabaseKerasSelenium Jan 2024 - May 2024 Data Science Researcher Cisco, West Lafayette, IN\nBuilt demand forecasts for 15 product lines with XGBoost and Random Forest (about 90% accuracy). Made Plotly and Dash dashboards that saved about five hours of reporting a week.\nXGBoostRandom ForestPlotlyDash Jun 2022 - Aug 2022 Web Development Intern Leadzen.ai, Mumbai, India\nHelped redesign the company website. Mostly JavaScript and CSS on the pricing pages.\nJavaScriptCSSElementor ","externalUrl":null,"permalink":"/PersonalWebsite/experience/","section":"Akhil Kota","summary":"Research labs, startups, and a few industry projects.\nFeb 2026 - Present Machine Learning Researcher Li Group, Prof. Can Li, Purdue ChemE\nWorking on neural methods that use difference-of-convex programming for hard optimization problems. I work on the PyTorch training code, including custom autograd, losses, and how we pick anchors. Goal is a conference paper.\n","title":"Experience","type":"page"},{"content":"A selection of things I’ve built—from study tools and competition platforms to computer-vision experiments and Discord bots.\n01 Project TickStudy A full-stack study companion that brings timers, task lists, habit tracking, and study statistics into one place. It includes both regular and Pomodoro sessions, so progress is easier to see and keep up with.\nReactTailwind CSSNode.jsExpressPostgreSQL View on GitHub ↗ 02 Project TurnMetric A native iOS speedcubing timer built for focused, private practice. It handles WCA-style timing, offline scrambles, per-event history, penalties, and statistics entirely on-device—without accounts, ads, or cloud sync.\nSwiftUISwiftDataSwift Concurrencycubing.js View on GitHub ↗ 03 Project cubecomp An online platform for running Rubik’s Cube competitions. Organizers can create public or private meets, run WCA-format events or head-to-head brackets, publish scrambles, and track results on live leaderboards.\nNext.jsSupabaseTailwind CSScubing.js Live demo ↗ View on GitHub ↗ 04 Project ClassClear An academic planning tool that turns messy course syllabi into a clearer timeline. Its dashboard uses AI to pull out important dates and requirements, giving students one organized place to see what is coming up.\nReactNode.jsExpressPostgreSQLOpenAI API View on GitHub ↗ 05 Project gears-bot A modular Discord bot built for Engineers for Sustainable World communities. Alongside moderation and server utilities, it supports reaction roles, reminders, XP and ranks, project lookups, calendar and social integrations, and chapter-specific tools.\nTypeScriptdiscord.jsSupabaseVitest View on GitHub ↗ 06 Project compvision-timer A browser-based Rubik’s Cube timer that uses webcam hand tracking instead of a physical stackmat. It recognizes intentional open-hand placement, backdates confirmed events to reduce debounce lag, and keeps solve history and averages in the browser.\nTypeScriptReactMediaPipeVitelocalStorage Live demo ↗ View on GitHub ↗ 07 Project CardQuest A Java flashcard app for making and studying custom decks. The goal is simple: make it easy to create cards and turn them into a more focused, repeatable study session.\nJavaFlashcardsStudy tool View on GitHub ↗ 08 Project Productivity-Bot A Discord bot that brings lightweight study and task planning into the conversation. It offers Pomodoro sessions, study-time leaderboards, task creation and completion, and a little motivation when you need it.\nPythondiscord.pyMongoDBAsyncIO View on GitHub ↗ 09 Project BoilerScout Natural-language event discovery for the Purdue campus. Students can search talks, free food, clubs, and more with Typesense hybrid search, campus geo filters, and RouteScout walking-corridor discovery between buildings.\nNext.jsTypeScriptTypesenseLeafletTailwind CSS View on GitHub ↗ More experiments and unfinished ideas live on GitHub. Browse all repositories ","externalUrl":null,"permalink":"/PersonalWebsite/projects/","section":"Akhil Kota","summary":"A selection of things I’ve built—from study tools and competition platforms to computer-vision experiments and Discord bots.\n01 Project TickStudy A full-stack study companion that brings timers, task lists, habit tracking, and study statistics into one place. It includes both regular and Pomodoro sessions, so progress is easier to see and keep up with.\n","title":"Projects","type":"page"},{"content":"A few ML projects from labs and industry.\nHard nonconvex optimization with neural nets\nDifference-of-convex neural optimization Li Group, Purdue ChemE, Feb 2026 - Present\nModels that use DC programming inside the network, with custom PyTorch autograd and better anchor selection.\nTraining setup works and early results look good. Working toward a conference submission.\nPyTorchDC programmingCVXPYOptimization See related experience Cell segmentation and tracking in noisy video\nMicroscopy and microplastics Prabhakar Group, Purdue, Sep 2025 - May 2026\nCellpose and TrackMate pipeline, fine-tuned on our labels, plus Python tools for cytotoxicity analysis.\nFalse-positive microplastic detections dropped by up to 21%. OUR Distinction List and an AAAR submission.\nCellposeTrackMateOpenCVMicroscopy See related experience Demand forecasts by product line\nProduct demand forecasting Cisco, Jan 2024 - May 2024\nXGBoost and Random Forest models on 15 product lines, plus Plotly and Dash dashboards.\nAbout 90% forecast accuracy. Saved roughly five hours of reporting each week.\nXGBoostRandom ForestPlotlyDash See related experience Supported by Purdue OUR and SURF. Email me if you want to work together.\nFull experience Get in touch ","externalUrl":null,"permalink":"/PersonalWebsite/research/","section":"Akhil Kota","summary":"A few ML projects from labs and industry.\nHard nonconvex optimization with neural nets\nDifference-of-convex neural optimization Li Group, Purdue ChemE, Feb 2026 - Present\nModels that use DC programming inside the network, with custom PyTorch autograd and better anchor selection.\n","title":"Research","type":"page"},{"content":" ","externalUrl":null,"permalink":"/PersonalWebsite/resume/","section":"Akhil Kota","summary":" ","title":"Resume","type":"page"},{"content":"","externalUrl":null,"permalink":"/PersonalWebsite/series/","section":"Series","summary":"","title":"Series","type":"series"},{"content":"What I use most often.\nLanguages Python, Java, C, C\u0026#43;\u0026#43;, JavaScript, TypeScript, SQL, HTML, CSS\nMachine Learning PyTorch, TensorFlow, Keras, scikit-learn, XGBoost, Cellpose, CVXPY, Pandas, NumPy, OpenCV, Matplotlib, Seaborn, Plotly, Dash, Streamlit\nWeb / Apps React, React Native, Node.js, Express.js, Flask, REST APIs, OpenAI API\nData / Cloud PostgreSQL, Supabase, Firebase, Google Cloud Platform, Docker, GitHub Actions\n","externalUrl":null,"permalink":"/PersonalWebsite/skills/","section":"Akhil Kota","summary":"What I use most often.\nLanguages Python, Java, C, C++, JavaScript, TypeScript, SQL, HTML, CSS\nMachine Learning PyTorch, TensorFlow, Keras, scikit-learn, XGBoost, Cellpose, CVXPY, Pandas, NumPy, OpenCV, Matplotlib, Seaborn, Plotly, Dash, Streamlit\nWeb / Apps React, React Native, Node.js, Express.js, Flask, REST APIs, OpenAI API\n","title":"Skills","type":"page"},{"content":"","externalUrl":null,"permalink":"/PersonalWebsite/tags/","section":"Tags","summary":"","title":"Tags","type":"tags"}]