Narek Maloyan


AI Research Engineer & PhD Candidate

Building AI coding assistants at Zencoder and researching AI safety — jailbreaks, prompt injections, and trojan detection in LLMs. PhD candidate at Lomonosov Moscow State University. Teaching Deep Learning at MSU since 2021.


Experience

Zencoder logo
Zencoder — AI Research Engineer (2024–present) Building AI-powered coding agents. Led team to #1 on SWE-bench Verified (May 2025).
Viasat logo
Viasat — ML Engineer (2021–2024) Video highlights extraction, movie recommendations, semantic search optimization.
TrendMD logo
TrendMD — ML Engineer (2019–2021) Medical article recommendations, speech-to-text for doctors, uplift modeling.
Burdenko NSI logo
Burdenko NSI — ML Engineer (2020–2022) MRI analysis for brain tumor classification, stroke risk prediction.
ODS logo
ODS.ai — Community Contributor (2020–2022) Data science competition pipelines, organized ML competitions with 1000+ participants.

Teaching

Deep Learning — Moscow State University (since 2021)

Graduate-level course covering neural architectures, optimization, and practical applications.

AI & ML Cheatsheet

Open-source reference guide covering key concepts across ML and data science.

manim-js (339+ stars)

TypeScript port of 3Blue1Brown's Manim for creating math animations on the web.

Watch on YouTube


Selected Publications

Full list on Google Scholar


Blog

Prompt Injection Attacks on Agentic Coding Assistants 2026
Breaking the Protocol: MCP Security Analysis 2026
The $20K Bug That Changed How We Think About Evals Mar 2026
Zencoder Leads SWE-bench Verified with 70% Success Rate May 2025
LLM-as-a-Judge Vulnerability to Prompt Injection 2025
Adversarial Attacks on LLM-as-a-Judge Systems 2025
Prompt Injection Attacks in Defended Systems 2024
Trojan Detection in LLMs 2024

All posts


Contact

Available for full-time roles, research collaborations, consulting, and speaking engagements.