Almas Perneshev avatar

Hi, I’m Almas,
a Software, Security & ML/DL Engineer and Math Enthusiast

I'm always eager to learn new technologies and utilize them in my problem-solving approaches. I am fluent in C/C++, Java, Python, JavaScript, TypeScript, SML, and x86 Assembly. My fields of Interest are Full-Stack Engineering and areas related to Big Data, Deep Learning, Privacy/Security, and Natural Language Processing.

Whenever possible, I apply my passion in teaching Computer Science and Mathematics to graduate, undergraduate and high-school students, where I have already instructed over 2000 students.

Currently, I'm concerned about the future of Security and Privacy of AI-Systems. 

Work Exprerience

Lead Software Engineer

  • RazeMath | January 2026 - April 2026 | Part-Time

As a lead engineer, I managed a team of 5 engineers, reviewing Pull Requests, providing actionable code feedback, and planning sprints to keep delivery on schedule. Owned the development and monitoring of infrastructure and its observability to maintain the system’s availability, and contributed to key product decisions alongside leadership.

Full Stack Engineer

  • RazeMath | June 2025 - December 2025 | Part-Time

As a full-stack engineer at a start-up, I’m developing a next-generation AI tutoring service to integrate with universities, where I worked on transitioning from Streamlit to Next.js, designing instructor’s Dashboard, setting up SSO Auth, and enhancing the prompt page using Next.js, Firebase, Django Rest, Langchain, Clerk and Vercel

Tutor: Computer Science and Mathematics

  • CalPoly SLO | Fall 2022 - Spring 2025

Tutored over 1000 students on courses: Calculus I - IV, Linear Algebra, Discrete Mathematics, Data Structures & Algorithms, Programming in C/Java/Python, Theory of Computation, Database Systems, Distributed Computing, Design & Analysis of Algorithms, Systems Programming, Cybersecurity, Software Engineering

Teacher Assistant: Theory of Computation (CSC445)

  • Cal Poly SLO | Fall 2024

Conducted supplemental sessions, exam preparation sessions, office hours, and assisted with grading over 1000 students for the selected courses.

AI/LLM Code Quality Tester

  • Outlier | October 2023 - Present

Tested LLM-generated code responses across multiple programming languages, queries, and frameworks such as Python, Java, C/C++, JavaScript, MySQL, ExpressJs, and ReactJs. Authored detailed suggestions and fixed not valid generated code responses to enhance performance, security, and scalability of LLM generated response.

SI Leader (TA): Programming in C/C++ (CS50/CS52)

  • Santa Monica College | Spring 2022

Conducted supplemental sessions, exam preparation sessions, office hours, and assisted with grading over 1000 students for the selected courses.

SI Leader (TA): Multivariable Calculus (Math 11)

  • Santa Monica College | Spring 2020 - Spring 2022

Conducted supplemental sessions, exam preparation sessions, office hours, and assisted with grading over 1000 students for the selected courses.

SAT Math Instructor

  • Grantly | Sep. 2021 - Feb. 2022

Led SAT Math preparation sessions for 4 groups of 5-10 students, focusing on critical math topics. Developed and administered a structured syllabus, session agendas, created study materials and mock exams to simulate real test conditions. Achievements: My students averaged 695/800 - 174 points above the national average in 2022. Personally scored 780/800 in Math, ranking in the World’s Top 1% in 2017.

Projects

...

[Team Project] Donkey Kong Reinforcment Q-Learning

  • 2025

This project is an implementation of a reinforcement learning agent for Donkey Kong, developed as part of CSC 480 at California Polytechnic State University.

...

[Personal Project] poly/Place

  • 2025

React, NodeJs, Supabase (Postgres), JWT, Websockets, Azure, Vercel

...

[Open Source] Potions Game

  • 2025

TypeScript, Ably, Superbase (Postgres), SQLite3, CI/CD, Docker, Terraform, Prometheus

...

[Paper] Privacy-Preserving Machine Unlearning: An Empirical Study of Selective Synaptic Dampening Against Membership Inference

  • 2025

Python, PyTorch, CNN, VGG, Membership Inference Attack , Selective Synaptic Dampending

...

[Team Project] Audio Music Genre Classification

  • 2025

Python, PyTorch. CNN, VIT, VGG16, MelSpectrogram, Data Augmentation