Engineer
Google Summer of Code Contributor 2023, 2024 and mentor 2025 | Backend Developer · Chennai · writing since 6 Jun
Day 22 · public proof
Their logs18
▶ Played as a game · Pick & Defend
E
Engineer 2AD4Google Summer of Code Contributor 2023, 2024 and mentor 2025 | Backend DeveloperDecided · 27 Jun
I usually follow this way when starting fresh on anything.
Big Picture -> A high level view or a short video or a tutorial showing the full process, what happens when they run a certain piece of code, what they do, just sit, hear and concentrate.
Build -> Now, the most efficient way of learning actually starts here, when I start to build a project, I have a good understanding of what I am going to do and not just following a tutorial, but this also becomes asking "what if's? and whys?" as we build it slow, I switch to deep dive modes occasionally.
Deep dive: Sometimes, a 10 min tutorial portion here ends up taking over an hour fighting with different variations of the code. And its best to do it now than "let's figure out later", because obviously I know (I never will) if I let that go now.
So doing this, makes a 1 day learning plan to 3 day plan unexpectedly, but worth it.
The first step seems necessary for me: Usually I compare it with math class in 7th grade (I guess) when teacher introduced us to "Trigonometry" -> I didn't like it, because when I ask what does this tan, sine, cos mean, the answers are "Formula" - ok but why? And when I finally realized in grade 10 these are actually pretty cool stuff used to figure out investigations where we can reveal a lot of info with known data with these concepts, now things clicked. I think for every concept those clicking moments of "why are we learning this" is important and that makes thing far more easier to understand. · I'll know I was right in 2 weeks if: It had worked out so far.
Free write20 Jun 2026
Back to basics and nuances on LLD, Java and Concurrency.
Things are straightforward to implement until one hears "What if two try to modify things at the same time?" - What if 2 requests are made at the same time and in the case of RateLimiters where we decide what if 2 requests are made at the same time. A concurrent datastructure helps, but when intially the structure is null, and we let multiple requests in? that now requires to synchronize the logic block as well on the specific user data. So trying to go a bit deep questioning and answering, and every time a new answer comes up.
Free write19 Jun 2026
Today was spent most on polishing the basics and preparing for the typical things on the side of LLD. Thought I would follow some tutorial to start off, but every single tutorial had the same pattern and pre-filled code. So took a high-level look and started code than binge watch. A 1 hour tutorial takes 4 hours to code if we actually start to question decisions, but much more interesting to solve and as well quite challenging to think other ways since the tutorials just etch a structure hard to let the brain think otherwise. Also had a communication assesment today with a financial company, seems like few companies do these assesments, non-tech but interesting. Plans for tomorrow: Might finish the different versions of limiters and also go back to the RAG. Oh, I think I also learned something nice that helps with RAG accuracies, so learning that is going to be interesting as time allows for tomorrow.
Built18 Jun 2026
I built the Section Chunking
Wonder18 Jun 2026
I wonder few decisions on choosing skills over yoe
Decided17 Jun 2026
I picked building my own LLM Judge to evaluate over available forks because Java's AI APIs are not as rich as Python's
Java's AI APIs are not as rich as Python's
Free write17 Jun 2026
Technical documentation over anything
Right when I was digging into RAG, thought why not take a look at Spring AI documentation, may be that might have something related to Evaluation. And once one starts to read and understand technical documentations, no LLM can stand to its beauty. Yeah today LLMs have instant answers to every single question but they are contained and we close the loop, docs feel way more than that. Though reading big docs are not so easy to understand and complete, Spring AI docs feel quite the starting story when things are still in the development. I might not end be able to read it in full while grabbing every detail but still even a single page is going to teach a lot more than a structured tutorial.
Decided16 Jun 2026
I picked section chunking over semantic chunking strategy because of computation and project requirement
of computation and project requirement
Figuring out15 Jun 2026
I want to figure out Accuracy of RAG pipelines
Stuck14 Jun 2026
I'm stuck on Leetcode Weekly 506
Built13 Jun 2026
I built RAG pipeline for the mert knowledge base.
Free write12 Jun 2026
Transitioning from raw RAG to Spring AI built in ways. Just starting off but that just shows
1. How easy these frameworks make it handling the complexities
2. But only with the understanding of the under the hood it really feels "Woah"
Learned11 Jun 2026
Continuing on the RAG series from mrdbourke
Free write10 Jun 2026
More on RAG and embedding the old way without Lang chain, so the use of lang chain will actually make sense later.
Free write9 Jun 2026
Slightly more into LLM (not much) but more time on networking. And setup to understand testing practices and codeRabbit workflow
Learned8 Jun 2026
LLM from Andrej Karpathy
Free write7 Jun 2026
Upsolving and Understanding Dynamic Programming problem from Leetcode 184 biweekly
Learned6 Jun 2026
RAG architecture from Krish Naik
How Engineer thinks
Decides and owns the call — moves before being told.
↳the log that shows itarticulate·curious·decisive
◈ Open to being sharpened
How Engineer thinks, mapped · lit by their own logs · 6 showing
How they reason · the four moves
Evidence
argue from what you saw, not what you assumed
Alternative
name the option you didn't take
Tradeoff
state the cost of the call
How they work · the qualities founders read for
Curious
you ask, not just assert
Articulate
you write the full why
Ownership
you own the call & the miss
Showing — it’s in their logs; tap to see them allno %, no rank — evidence, not a score.
Milestones2 earned
Day 20 · Jun 25Day 10 · Jun 15
About
JavaPostgresSpring BootCi CdGitGithubHtmlObject Oriented ProgrammingAgile WorkflowCode ReviewAndroidData engineering
Want a page like this?
Write what you tried, a few lines a day. People who’d hire you read it.
Start your Proof