Software Engineer, Driven by curiosity || @web3bridge cohort XI, I enjoy writing linktr.ee/devvmichael

root
I am grateful for the present and excited for the future 😄
1
2
16
618
📍
Often creating something simple is more difficult than creating something complex…
1
35
Ultimately…
Knowing how is good, knowing why is better..
1
42
A M D retweeted
You learn by doing...
1
6
57
A M D retweeted
There’s a big difference between trying to win and trying not to lose.
1
1
27
Day 10/180 🦀 Finished my Free List Allocator in Rust. Today was about putting the pieces together: allocation, splitting, deallocation, reuse, and coalescing adjacent free blocks. Implementation: github.com/michojekunle/rust…
Day 9/180 🦀 Today I started building a FreeListAllocator in Rust. I went from understanding why a bump allocator can't easily reuse freed memory to learning how a free list tracks available regions using intrusive metadata stored directly inside the free memory.
1
5
97
Next up: • Integrate it with a small binary using #[global_allocator] • Benchmark it against the system allocator with Criterion • Document its limitations • Deep dive into UnsafeCell and raw pointers Still a lot to learn about memory management. We go again tomorrow. 🦀
1
4
17
Spoken Word Poetry at UNGA A room filled with World Leaders, Heads of State & Entrepreneurs. The theme: “The Youth Dividend: Turning AI & Green Investments into Skilling & Livelihood Opportunities for Youth.” 📍 UNICEF HQ, New York 🇳🇬✈️🇺🇸 #UNGA81 #SpokenWord #UNICEF #GenU
38
141
1,013
12,857
“There is healing in Your Name and it’s Your healing we proclaim” #HealingHands #FOAW #greatmantakit #flamesofawildfire
10
68
281
2,761
Day 9/180 🦀 Today I started building a FreeListAllocator in Rust. I went from understanding why a bump allocator can't easily reuse freed memory to learning how a free list tracks available regions using intrusive metadata stored directly inside the free memory.
Day 8/180 🦀 Lightweight return to Rust. Completed a simple quicksort algorithm in Rust, then wrote a parallel quicksort algorithm as well. Benched both and got to see the trade-offs. The parallel quicksort was better at larger inputs, about 3x faster on a million inputs.
1
5
244
I implemented the FreeListAllocator from scratch and added tests for the allocator initialization. Code + notes: github.com/michojekunle/rust… More allocator work tomorrow.
3
16
A M D retweeted
Watch full.message via Youtube on Pastor Poju Oyemade Channel. Title Meditation and the law of growth. Convert the groaning and travail within into products and services through which creation will experience our liberty in Christ
9
134
703
18,417
A M D retweeted
The disappointment people feel atimes may be because they had no respect for the process. They want an instateanous manifestation and results in something that required a process for fruition. Their words and steps initiated a real process but they were expecting the spectacular rather than the supernatural. When they didnt get the result in the very short time frame they gave themselves and God, in disappointment, they withdrew and shut down the process which had already began. Sometimes play the long game. It is through faith and patience we inherit and not faith and nastiness.
19
167
756
25,657
Been studying system design, Rust, and DSAR, and one thing I learned recently was the trade-off between arrays and linked lists. Arrays are great for reads, while linked lists are better suited for writes. I initially wondered if production systems combine both...🧵👇
Day 8/180 🦀 Lightweight return to Rust. Completed a simple quicksort algorithm in Rust, then wrote a parallel quicksort algorithm as well. Benched both and got to see the trade-offs. The parallel quicksort was better at larger inputs, about 3x faster on a million inputs.
1
6
197
But that gets complex. Instead, systems like RocksDB and Cassandra use LSM trees, which are designed around efficient writes while using compaction strategies to balance read performance. Another rabbit hole unlocked. 🦀 Learn more about LSM trees here: piped.video/I6jB0nM9SKU?si=cpPT…
2
17
Ethereum just made the direction clear. The @ethereumfndn is targeting a quantum-resistant Ethereum L1 across execution, consensus and data by December 2029. And this is exactly why we started building @geanclient at @geanlabs . GEAN is our independent Go implementation of Lean Ethereum consensus, built from Africa, focused on post-quantum cryptography, interoperability and client diversity. A few months ago, this looked like long-term research. Today, the EF says post-quantum work is moving into near-fork delivery. The next few years will require more client teams, more testing, more cryptographers and more independent implementations. We’re already in the arena. Building. Breaking. Testing. Interoperating. Not just building on Ethereum from Africa. Helping build Ethereum itself. GEAN 🇳🇬🌍 Read full news article: blog.ethereum.org/2026/09/07…
5
23
354
Day 8/180 🦀 Lightweight return to Rust. Completed a simple quicksort algorithm in Rust, then wrote a parallel quicksort algorithm as well. Benched both and got to see the trade-offs. The parallel quicksort was better at larger inputs, about 3x faster on a million inputs.
Day 7/180 🦀 Review day. This week was mostly about getting back into Rust properly, going from basic project structure and data structures to actually implementing, testing, and benchmarking things. A lot of basics I haven't touched in a while came back. 1/6 🧵
1
5
480
So every time, the left subarray had significantly more work to do, and over a larger number of inputs, that affected it so much that at a million inputs I got a stack overflow because the number of recursive calls exceeded the stack. Good lesson tho.
1
12
I changed the pivot selection to the middle element, and the results changed significantly. Small decisions in code design can create large overheads. Think about edge cases, and test and benchmark rigorously for scale. We go again today.
7