Skypebakyt.kosjan

Emailbakytkh@outlook.com

Backend/NLP:
Node.js, Python, Pytorch

Developing Node-Python backend apps
ML (NLP) on Pytorch and Sklearn
Data science approach to business problems
SEE PORTFOLIO

MY STACK

BACKEND

On the backend, I have worked with Node.js & Python for prod apps. In databases, I have familiarity with SQL, MongoDB and Redis. For deployment, used AWS, flyctl, nginx-linux

MACHINE LEARNING

Pytorch and Sklearn are two main prototyping frameworks for models, which can be deployed on-premise with Flask, uWSGI or via AWS SageMaker cloud app

FRONT-END

Limited exposure to Vue.js on the front-side. Relying on Bootstrap-vue for component library, I built a couple of web apps outlined below

WORK EXPERIENCE

Backend developer
(Sep, 2021 - Present)

Have worked in a Swedish startup through Proxify platform to develop serverless functions for the warehouse logistics project. Database was Postgres and the backend stack was AWS Lambda-SQS-SNS. Wrote 3-4 API functions from scratch on Typescript and Javascript

Web developer
(Jan, 2020 - Aug, 2021)

Developed a full-stack project on Node.js-Vue.js, which as a pizzeria website. The cart of items was stored in Redis, with data persistence in Mongo. Secondly, developed messenger bots on Facebook and Telegram, collecting order for barbers and bookings for restaurants

Data analyst
(Jul, 2018 - Nov, 2019)

Working in post-sales retail credit department, I built customer classification and segmentation models on Python. Data was pulled from Oracle SQL, while pre-processing was in Pandas. The modelling toolkit was Sklearn, and data visualization via Seaborn

PORTFOLIO

team img 2

Chatbots

DialoGPT

ML app, calling DialoGPT via Python package, to render conversation bot, which can be trained on a custom text. A basic server is on Express and an SPA with Vue.js. More info here

team img 3

Name-guess

Pytorch

Character-level RESNET, predicting the origins (country) of a given name based, trained on corpus of name-vs-country dataset. More info here

team img 3

Pizzeria

Full-stack app

A pizzeria website on Node.js-Vue.js, with front for clients to make orders and dashboard for admins to add/edit/delete items. Dbs are Redis and Mongo. More info is here

team img 1

Movie-CRUD

Express-Mongo

This repo includes REST API on Express, saving movie data into Mongo with OMDB metadata. App containerized with Docker. Testing with Github actions. More info here

EDUCATION

www.isb.edu

Indian Business School
Mohali, India (2017 - 2018)

Post-graduate program in Management
Main courses: Management, Marketing, Entrepreneurship

www.graduateinstitute.ch

Grad. School of Inter'l Studies
Geneva, Switzerland (2003 - 2005)

Master program in History
Main courses: History, Economics, Econometrics

www.metu.edu.tr

Middle East Technical Univeristy
Ankara, Turkey (1999 - 2003)

Bachelor of Arts in International Relations
Main courses: Politics, Economics, Calculus

SELECTED COURSERA COURSES

portfolio img 1

Machine learning and Data analytics

Moscow Institute of Physics and Technology, Yandex

VIEW LINK
portfolio img 2

Deep Neural Networks with PyTorch

IBM

VIEW LINK
portfolio img 3

Introduction to TensorFlow

Google Cloud

VIEW LINK
portfolio img 4

Fundamentals of Reinforcement Learning

University of Alberta, Alberta Machine Intelligence Institute

VIEW LINK
portfolio img 3

Dive in Python

Moscow Institute of Physics and Technology, Mail.Ru Group

VIEW LINK
portfolio img 4

C Programming: Pointers, Arrays, and Recursion

Duke University

VIEW LINK
portfolio img 1

Foundations of marketing analytics

ESSEC Business School

VIEW LINK
portfolio img 2

Programming with Google Go

University of California, Irvine

VIEW LINK

CONTACT INFO

ADDRESS

The preferred channel of communication is email

bakytkh@outlook.com

Valetta, Malta