About
Salih Keskin
I am Salih, a Literature graduate from (American Culture and Literature) Istanbul University, where we focused more on critical thinking on everything involving a letter, rather than just "studying" literary works in what may be perceived as "traditional" sense.
During my college years, for unknown reasons, I developed a fondness for programming, which in my opinion is just another form of literature, unrelated to maths unless you are a computer doing the calculations. What started as curiosity evolved into a passion that eventually overshadowed my academic field.
While my formal education is in literature, I've found that the analytical thinking and problem-solving approach I learned there transfers surprisingly well to programming. My literature background taught me to deconstruct complex texts and ideas - a skill that's proven valuable when breaking down complex programming problems.
As I did programming with numerous tools for years (since 2020-2021 I regularly write programs for myself) as a hobby regularly; I can safely say my computing&engineering&programming skills have exceeded my literary skills, and that if given the necessary documents, I can accomplish anything given&asked at this point, but it'll take time if I have little to no experience with them! (I truly think programming is just building up experience at this point, you cant just know a programming language!)
I've embraced this unconventional path from humanities to technology, and while I may not be utilizing my literature degree in the traditional sense, the critical thinking and communication skills it provided have been invaluable in my tech journey. My interdisciplinary background gives me a unique perspective when approaching technical challenges.
My favorite productive hobby -other than general purpose programming for my needs&solving my own problems- is Large Language Models and I follow the developments since I discovered sentence transformers in late summer 2022 (I was pretty late), I have been doing my experiments too since then (such as by fine-tuning and pre-training experiments of my own)
Skills
Large Language Models (Local)
I run various open-source language models locally on my personal machine, and closely follow the development in this field. I test models myself, and I've experimented with fine-tuning and quantization techniques myself at various points.
Python
The primary programming language I use for most of my projects. I've used it extensively for web development, data processing, automation, and AI/ML applications.
GDScript
A Python-like language used in the Godot game engine. As a result, I have experience with this language ranging back to 2021.
Flask
Python web framework I use for creating web applications and APIs. This website is built with Flask!
Vibe Coding
The art of creating software by describing ideas to AI coding assistants. I've embraced this modern approach to development that became recognized in 2025, which allows me to rapidly prototype ideas without writing every line of code manually. While I understand the code being generated (and most of the time, create more efficent algorithms than the ones generated), I find this approach efficient as it saves a lot of time and enchances creativity alongside giving me more creative freedom.
Critical Thinking
This was the main thing that was taught in our lectures, so I am presuming I am experienced enough.
English
I literally studied American Culture and Literature, so this should be obvious.
Curiosity
I have listed this as a skill because most of skills listed here are a result of this, in both positive and negative ways.
Creative Artistic Thinking
I try using new technologies in a creative artistic way, trying to figure out where would they look best, used and utilized, given time. My understanding of local LLM is mainly driven by this factor for instance.
Other Skills
C
My first programming language where I learned basic programming logic, though I quickly moved to higher-level languages after grasping the fundamentals.
C#
Used primarily in Godot projects when GDScript wasn't sufficient. Limited experience focused on game development scenarios.
Java
Used briefly for one game development project. While I understand the language, I prefer other languages for most of my work.
Nuxt
I used to use Nuxt before I noticed it caused rendering problems on other people's machines, so I switched to Flask. There was also a frontend&backend divide that caused issues that flask unified.
Vue
Vue empowers Nuxt, therefore yes, I have used it.
Japanese
I study Japanese independently, having restarted after pausing at the elementary (N4) level. This study also includes a focus on Classical Japanese (Kobun) and Kanbun Kundoku, as I find this parallel approach deepens my understanding of the modern language's structure. Therefore; I think I know a bit or two about this language
OpenAI
I used OpenAI API extensively in closed and open projects of mine (from normal models to embeddings) as it is standard in many providers (LM Studio, Ollama, OpenAI itself, OpenRouter, etc.).
Experience
AI Coach&Software Engineer
Here, I attempted to move the company itself into using local LLM solutions, and in my opinion, I did my best for 3 months.
However, the company itself didn't think I was doing enough to contribute to the company, (I was given unrelated tasks that I obviously didnt excell at (such as automation using hardcoded processes), and I did most I could using local LLM instead) so I resigned after mutual inability to communicate, but most importantly, the fact that I wasn't learning anything new and improving myself was a major factor for this.
Education
Bachelor Of Arts in Humanities (American Culture and Literature)
Certifications
Pedagojik Formasyon (Teaching Certification)
Completed an additional year of pedagogical training that qualifies graduates from non-education faculties to become eligible to teach in schools. This certification program extended my academic studies from 4 to 5 years.