Victor Nwachukwu

[Victor Nwachukwu @DVectorPrime]

The frustrating part? After all that logical feature engineering, my XGBoost ROC-AUC score is still stuck at 0.65. 😅 I’m heading back to the EDA and visualization phase to hunt for deeper patterns.

Data Scientists: where do you look when feature engineering doesn't work? 👇📉

[Victor Nwachukwu @DVectorPrime]

Back on my ML project, LendIQ, I caught a major real-world flaw: absolute currency values. The dataset is in USD, but my local currency is Naira. A 500k loan means wildly different things. To make the model currency-agnostic, I engineered Loan-to-Income ratios instead. 🌍💵

[Victor Nwachukwu @DVectorPrime]

Took on LeetCode 2415 (Reverse Odd Levels of Binary Tree) using BFS. Worried my nested loops were O(n^2), I put on my CS student hat and calculated the exact running time function. Fun realization: Python's deque is a doubly linked list, so index lookups aren't true O(1)! 🌳⏱️

[Victor Nwachukwu @DVectorPrime]

Now for the final boss: the ML tier. My XGBoost model is currently performing slightly better than a coin toss 🪙. The culprit? A highly imbalanced dataset tanking the ROC-AUC score.

ML Engineers (and @AnthropicAI folks!)—what’s your go-to strategy for imbalanced data? 👇📉

[Victor Nwachukwu @DVectorPrime]

Been a heavy coding week! 🚀

First up: DSA. Tackled "Kth Smallest Element in a BST". The concept was easy, but translating it to code and managing the call stack took serious troubleshooting. Finally seeing that green "Accepted" was worth the grind. 🌳💻

[Victor Nwachukwu @DVectorPrime]

Next, full-stack work on LendIQ (my loan prediction app). Built the UI and wired it to the backend. Because I planned the architecture earlier, tweaking the backend on the fly was stress-free. There is nothing like seeing a full system click together! ⚙️🔧

[Victor Nwachukwu @DVectorPrime]

But the data is tough. I'm dealing with a highly imbalanced dataset and my ROC-AUC is currently sitting at 0.654. I want to push it to at least 0.75 before deployment.

Data scientists & ML engineers: What are your go-to strategies for improving ROC-AUC on stubborn models?

[Victor Nwachukwu @DVectorPrime]

Jumped back into Lendiq, my 3-tier loan prediction app. Since it's my first time using Prisma & MongoDB, I felt completely lost opening the codebase. What saved me? A step-by-step "Design Game Plan" I documented earlier.

Lesson: Document your architecture. It’s a lifesaver! 👇

[Victor Nwachukwu @DVectorPrime]

Shifted to the ML tier to optimize my XGBoost model. To understand its decisions, I dove into SHAP. It took some workarounds, bypassing pipeline limits for scaled data & converting log odds to probabilities but finally seeing the feature impacts was worth it.

[Victor Nwachukwu @DVectorPrime]

A model needs a home! Putting my full-stack skills to work by spinning up MongoDB, navigating the Prisma ORM, and learning FastAPI to serve the model.

Also teaming up with @OritsesangbemiM and @davidchair3025 competition project. More updates coming, good to be back!

[Victor Nwachukwu @DVectorPrime]

Back from a posting hiatus! 🚀 Academics were heavy this semester, but the learning didn't stop, it just went under the radar. Let’s do a quick sprint through what I've been studying and building behind the scenes.

[Victor Nwachukwu @DVectorPrime]

Deep into Python ML. I'm building a Loan Default Prediction app and hit a wall: heavily imbalanced data. A model could approve everyone and hit 90% accuracy! It forced me to ditch basic accuracy and rely entirely on Precision, Recall, and ROC-AUC scores.

[Victor Nwachukwu @DVectorPrime]

The academic grind was real. Algorithms was a favorite, Assembly tested my patience, but Software Engineering completely rewired how I view the dev lifecycle. Also, connecting classroom Bayes' Theorem directly to the Naive Bayes algorithm in ML was a massive "aha" moment.

[Emmy @DevEmmy_]

I just shipped the tournament and in-game chatting feature. 🥹🥹

What do you think about this?
What feature would you like to see?

Here's the link - https://t.co/6jQ56DarPH

[Victor Nwachukwu @DVectorPrime]

Didn't get to code as much as I wanted, but still squeezed in some solid learning:

✅ Express.js: Writing to MongoDB
✅ ML: XGBoost pipelines & hyperparameter tuning ✅ Java: String methods (still fighting local vs instance variables tbh 😅)

[Hridoy Reh @hridoyreh]

📂 SaaS Stack
┃
┣ 📂 Frontend
┃ ┣ 📂 React
┃ ┣ 📂 NextJS
┃ ┣ 📂 Vue
┃ ┣ 📂 TailwindCSS
┃ ┗ 📂 Shadcn UI
┃
┣ 📂 Backend
┃ ┣ 📂 NodeJS
┃ ┣ 📂 Django
┃ ┣ 📂 Laravel
┃ ┣ 📂 FastAPI
┃ ┗ 📂 Express
┃
┣ 📂 Database
┃ ┣ 📂 PostgreSQL
┃ ┣ 📂 MySQL
┃ ┣ 📂 MongoDB
┃ ┣ 📂 Redis
┃ ┗ 📂 Supabase
┃
┣ 📂 Auth
┃ ┣ 📂 Clerk
┃ ┣ 📂 Auth0
┃ ┣ 📂 Firebase Auth
┃ ┣ 📂 Supabase Auth
┃ ┗ 📂 NextAuth
┃
┣ 📂 Payments
┃ ┣ 📂 Stripe
┃ ┣ 📂 Paddle
┃ ┣ 📂 Dodo Payments
┃ ┣ 📂 Lemon Squeezy
┃ ┗ 📂 Polar
┃
┣ 📂 Emails
┃ ┣ 📂 Resend
┃ ┣ 📂 SendGrid
┃ ┣ 📂 Mailgun
┃ ┣ 📂 Postmark
┃ ┗ 📂 Amazon SES
┃
┣ 📂 Storage
┃ ┣ 📂 AWS
┃ ┣ 📂 Cloudflare
┃ ┣ 📂 Google Cloud Storage
┃ ┣ 📂 Supabase Storage
┃ ┗ 📂 Uploadcare
┃
┣ 📂 Deployment
┃ ┣ 📂 Vercel
┃ ┣ 📂 Netlify
┃ ┣ 📂 Railway
┃ ┣ 📂 Render
┃ ┗ 📂 AWS
┃
┣ 📂 Domains and DNS
┃ ┣ 📂 Namecheap
┃ ┣ 📂 Hostinger
┃ ┣ 📂 Cloudflare DNS
┃ ┣ 📂 Google Domains
┃ ┗ 📂 SiteGround
┃
┣ 📂 Analytics
┃ ┣ 📂 Google Analytics
┃ ┣ 📂 Plausible
┃ ┣ 📂 PostHog
┃ ┣ 📂 Mixpanel
┃ ┗ 📂 DataFast
┃
┣ 📂 Monitoring
┃ ┣ 📂 Sentry
┃ ┣ 📂 LogRocket
┃ ┣ 📂 Datadog
┃ ┣ 📂 NewRelic
┃ ┗ 📂 UptimeRobot
┃
┣ 📂 DevOps
┃ ┣ 📂 Docker
┃ ┣ 📂 Kubernetes
┃ ┣ 📂 GitHub Actions
┃ ┣ 📂 CI CD
┃ ┗ 📂 Terraform
┃
┣ 📂 Search
┃ ┣ 📂 Algolia
┃ ┣ 📂 Meilisearch
┃ ┣ 📂 Elasticsearch
┃ ┣ 📂 Typesense
┃ ┗ 📂 OpenSearch
┃
┣ 📂 AI Integration
┃ ┣ 📂 OpenAI API
┃ ┣ 📂 Anthropic API
┃ ┣ 📂 Replicate
┃ ┣ 📂 HuggingFace
┃ ┗ 📂 Gemini API
┃
┣ 📂 Integrations
┃ ┣ 📂 Zapier
┃ ┣ 📂 Make
┃ ┣ 📂 n8n
┃ ┣ 📂 Pabbly
┃ ┗ 📂 Webhooks
┃
┣ 📂 Security
┃ ┣ 📂 SSL
┃ ┣ 📂 Cloudflare
┃ ┣ 📂 WAF
┃ ┣ 📂 Rate Limiting
┃ ┗ 📂 Secrets Management
┃
┣ 📂 Marketing
┃ ┣ 📂 Search Console
┃ ┣ 📂 Outrank
┃ ┣ 📂 Buffer
┃ ┣ 📂 Analytics
┃ ┗ 📂 Kit
┃
┗ 📂 Customer Support
┣ 📂 Intercom
┣ 📂 Crisp
┣ 📂 Zendesk
┣ 📂 Tawk
┗ 📂 HelpScout

[Victor Nwachukwu @DVectorPrime]

@saen_dev Thank you so much

[Victor Nwachukwu @DVectorPrime]

WE MADE IT! 90 unbroken days of code.

Started this challenge over the holiday to keep my motivation up (huge shoutout to my coursemate @DionDevs for the idea). Here is a look back at everything I built, broke, and learned over the last 3 months! 🧵👇

[Victor Nwachukwu @DVectorPrime]

💼 I am currently open to remote internships where I can bring my Full-Stack/ML skills to a team while balancing my degree.

Thank you to everyone who followed the journey! 🚀🔥

[Victor Nwachukwu @DVectorPrime]

As my UI coursework gets heavier, I'll be shifting from daily grinds to weekly updates.