jkru3.xyz

Recent Notes

See 82 more →

portfolio

6 min read

my aim is to build systems and infrastructure that can demystify the inner workings of the black box for ML models (MLOps). To the uninitiated, there is a perception that AI is a field reserved to pHd holders and the mathematical elite. I hope to contribute in shattering that illusion

note: my professional experience are deliberately excluded from this portfolio. This is a show case of my extracurricular learning and projects. If you want to know more about my work history, you can check out my LinkedIn or view my resume here

2025 learning objectives

 

my notes on these topics can be found at this repo

  • notes on system design
  • KodeKloud courses (start with DevOps path and Cloud first)
    • Kubernetes
    • GitOps
    • Go
    • CI/CD
    • Docker
    • Shell Scripting
    • Linux/Unix admin

MLOps

Kubernetes

  • set up a local Kubernetes cluster, deploy stateless applications
  • implement stateful workloads, PersistentVolumes, and StorageClasses
  • configure networking, ingress controllers, and create a demo project

GitOps

  • set up a local ArgoCD instance with minikube, deploy sample application
  • implement GitOps workflow for a personal project
  • build a CI/CD pipeline that integrates with your GitOps workflow

Go

  • build CLI tools with Go standard library
  • learn Go concurrency patterns (goroutines, channels)
  • build a DevOps utility tool in Go

Monitoring with Prometheus/Grafana

  • set up Prometheus locally, understand core concepts, implement basic exporters
  • set up Grafana dashboards with Prometheus and implement basic alerting

ongoing projects

 

Clearvote

Python Node.js TypeScript Flask React Next Docker AWS Lambda AWS Amplify AWS RDS AWS DynamoDB AWS IAM AWS Route 53 Mapbox OpenAI GeoPandas

 

  • app link
  • front-end repo
  • precinct-mapper repo (how we map user coordinates to geodata)
  • LLM pipeline and orchestrator repo (we can’t share this yet, sorry 🤫)  

provides over 3000 users with side-by-side comparisons of election candidates. Aims transcend party lines and provide prospective voters with a new way to visualize political intentions. Co-founded with Anaya Pandit

financial portfolio rebalancer

Android iOS Flutter Dart Hive C++ Python Docker Kubernetes

building an app that trains time-series models on stock data, which it can run through a simple mobile application to provide suggestions on how to trade as a long-term investment strategy. This is split into 3 parts:

  1. trade-dashboard: a front-end
  2. auto-trader: uses a simple EMA algorithm to predict which stocks to invest in
  3. time-series-forecast: TBD

Jobfindr

LaTeX Python

 

a crude tool i made to help me

  1. tailor resumes to specific job postings
  2. notify me whenever a company career page i’m interested in adds a new role  

right now, it’s finicky and only setup for personal use, but when i’ve refined it more, i’m interested in turning this into a simple open-source tool anyone can use

X/Twitter engagement maximizer

Python

a web scraper for twitter that logs into your account and figures out who has been engaging with your posts. It then creates a curated list of accounts consisting of only accounts who have engaged with you the most

jkru3.xyz

you’re looking at it (repo here)

previous projects

 

High Speed Rail project

D3.js Vega OpenAI

a data visualization project on the topic of High Speed Rail in the US. I was responsible for 100% of two of the visualizations

  • travel modes: is an intractable line graph made with D3 showing the coverage high-speed-rail provides when compared to car and air travel
  • opinion board: is a graphical representation of differing opinions on Reddit made with D3. I used GPT-4o to make a pipeline ingested web-scraped Reddit data, categorized the sentiments, and then decided on which sentiment was negative, neutral, and positive. This is an idea i’d like to expand on —being able to see how sentiment changes over time could be a really cool thing to visualize, especially for current events

i also contributed to the last visualization, which was originally made with Vega and adapted by my teammates

illustrAItor

Figma

a hackathon project i made with Elias Belzberg. We won

ProtoQA submission

PyTorch Hugging Face

an NLP research project submitted against the ProtoQA common-sense reasoning benchmark   Steve Harvey

everything always works out in the end

Spotify

 

i spent years of my life devoted to writing, playing, and performing in the musical arts. This is what I have to show for it  

future projects

 

projects take time from other things i enjoy so no guarantees i will have the time to start any of this. I have a a template for building profitable things that i use to determine whether or not something is worth working on

feel free to take any of this stuff and run with it, just make sure to credit me for the inspiration (better yet, reach out and let’s build it together)

Graph View

Backlinks