
Kima Rafayelyan
@kimarafayelyan
AI/ML engineer building banking voice AI with RAG, Armenian-language semantic search, and LLM inference optimization.
What I'm looking for
At CONVERSE BANK, I develop and optimize SQL queries, stored procedures, and scripts for financial datasets and business applications. I also automate financial reporting workflows and maintain database performance in a secure production environment.
I developed an Armenian banking voice assistant for questions about credits, deposits, and branch information. Its RAG pipeline uses FAISS semantic search and multilingual embeddings, with Gemini 2.5 Flash for response generation and LiveKit, Whisper, and OpenAI TTS for speech-to-speech interactions.
For my thesis, Optimizing LLM Inference with vLLM and LMCache, I evaluated inference systems through controlled experiments measuring latency, throughput, and GPU memory consumption. I also developed a LangGraph multi-agent research system that gathers, analyzes, validates, and synthesizes information into structured reports.
Experience
Work history, roles, and key accomplishments
Database & Software Developer
Converse Bank
Develop and optimize complex SQL queries, stored procedures, and scripts for large-scale financial datasets. Design and automate financial reporting workflows and maintain database performance in a secure production environment.
Education
Degrees, certifications, and relevant coursework
Yerevan State University
Bachelor's Degree, Informatics and Applied Mathematics
2022 - 2026
Activities and societies: Probability Theory & Mathematical Statistics, Linear Algebra & Mathematical Analysis, Data Structures & Algorithms, Object-Oriented Programming (C++), Data Analysis & Performance Metrics Evaluation, Large Language Models (LLMs) & Inference Optimization
Bachelor's degree in Informatics and Applied Mathematics with a thesis on optimizing LLM inference with vLLM and LMCache.
American University of Armenia
Specialized Course, Machine Learning
2025 - 2025
Specialized course in Machine Learning for Bio covering classification, regression, decision trees, dimensionality reduction, clustering, and neural networks.
Henk Foundation
Artificial Intelligence Program, Artificial Intelligence
2025 - 2025
Artificial Intelligence program covering Python for data science, machine learning fundamentals, and Linux command-line tools.
Availability
Location
Authorized to work in
Website
ysu.am/enSalary expectations
Job categories
Interested in hiring Kima?
You can contact Kima and 90k+ other talented remote workers on Himalayas.
Message KimaGet matched with your dream remote job
Sign up now and join over 250,000+ remote workers who receive personalized job alerts, curated job matches, and more for free!
