Contact
contact
Let’s connect.
Best way to reach me is email. For everything else, my profiles are one click away.
Email
abu2012only@gmail.com
Compose ->
GitHub
@Akuma277353
Open profile ->
LinkedIn
/in/abu-shaikh
Open profile ->
Tip: these are also in the Start menu.
Notes
Work in Progress
This website is still under development and is not yet finished. Some features may be incomplete or change over time.
This website is still under development and is not yet finished. Some features may be incomplete or change over time.
Click any folder to explore:
Quick vibe: Computing Science (AI) student curious about how software works end-to-end — from execution and data to real-world constraints.
Resume
Abubakar_Shaikh_Resume.pdf
PDF • …
Projects
Selected projects. “Info” gives a quick overview; “Open” takes you to the repo (if public).
Mini Compiler Back End: RISC-V to WASM
Private coursework: instruction decode/execution and a small CPU state model.
Pacific Atlas (Work in Progress)
WIP mapping project: interactive layers (quakes, volcanoes) + island explorer.
About Me
Hello, I’m Abubakar Shaikh.
How I think about building things
- Tracing behavior from input -> state -> output, including edge cases
- Understanding how programs execute and use memory
- Designing clear data models and querying them intentionally
- Building projects that are honest about real-world constraints
Beyond computer science
Outside of CS, I like staying active and hands-on. You’ll usually find me at a climbing gym, swimming, or folding origami when I want to slow down and think differently.
I’m also drawn to maps, visual design, photography, and learning about places and cultures. These interests influence how I approach software — I care about structure, flow, and how people actually experience what they’re using.
A few things about me
- I learn best by building, breaking, and fixing, especially tinkering
- I like understanding the mechanics, not just abstractions
- I value simple solutions that age well
- I’m curious by default, even outside my comfort zone
Skills
Programming languages
C
Python
Java
JavaScript
SQL
Bash
RISC-V Assembly
HTML / CSS
Tools & platforms
Git & GitHub
VS Code
Android Studio
Firebase
MongoDB
React (foundations)
CI / CD (intro)
Systems & execution
Memory & pointers
Program execution
Debugging
Performance basics
Data modeling
AI & data foundations
ML + Reinforcement Learning foundations. Comfortable working with data, evaluation, and clear model assumptions in coursework and projects.
NumPy
Pandas
Matplotlib
Neural Networks (intro)
Experience
Selected highlights
Hackathon project
AORS — Autonomous Overload Remediation System
Control systems • adaptive algorithms • real-time monitoring
- Built an autonomous closed-loop system that detects CPU overload on real hardware, applies throttling, and verifies recovery — mirroring AWS/GCP production patterns.
- Implemented temporal overload detection using sliding windows and multi-signal fusion (CPU trends, error rates, queue depth) with calculated confidence thresholds.
- Designed an adaptive work-intake controller using online learning (EWMA + Welford) to dynamically map request rate to CPU utilization.
- Trained a scikit-learn classifier to distinguish normal traffic spikes from retry storms, enabling targeted mitigation based on live telemetry.
- Shipped a FastAPI control plane with non-blocking telemetry and Streamlit dashboard — demonstrated recovery from 90%+ CPU saturation in under 30 seconds.
- Challenges: Tuning detection thresholds to avoid false positives, handling hardware variability across CPUs, and ensuring control loop stability under extreme load.
- Learnings: Control theory (feedback loops, stability, convergence), adaptive systems learning from live data, and production-grade patterns (separation of concerns, auditability).
Team project
AuroraChance
Applied software • consistency • edge cases
- Built an Android platform using probabilistic lottery sampling to fairly allocate high-demand events.
- Implemented real-time sync with Firebase Firestore across entrants, organizers, and admins under concurrent updates.
- Designed core data models (events, users, waitlists, notifications) to enforce predictable state and avoid race conditions.
- Integrated Google Maps SDK and Places API for location validation and map-based interaction.
- Collaborated via GitHub workflows and Agile Kanban to deliver 45+ user stories.
Personal project
Win95/98 Portfolio Website
UI polish • interaction design • vanilla JS
- Implemented a desktop-style window system with open, minimize, and close behavior tied to a taskbar.
- Built draggable and resizable windows with explicit z-index control and active-window tracking.
- Added non-modal info panels to avoid background dimming and scroll locking.
- Managed global UI state and event handling to keep interactions predictable.
Foundations
Systems & data
execution · debugging · data thinking
- Worked with low-level execution ideas (state, memory, and step-by-step correctness)
- Built comfort with data modeling + querying (relational concepts + practical DB workflows)
- Got used to validating results with small tests first, then edge cases
Foundations
ML / RL
evaluation · baselines · debugging
- Turned loosely defined problems into inputs, outputs, and evaluation metrics
- Implemented and debugged learning algorithms with careful training and evaluation checks
- Started from simple, explainable baselines before adding complexity
Info