Artem Meshkov
Senior Full-Stack & AI Engineer
I build fast, reliable and secure web applications
Projects
Projects that demonstrate range across frontend, backend, and infrastructure.
SPD Online
Real-time multiplayer roguelike dungeon crawler.
A real-time multiplayer roguelike dungeon crawler based on Shattered Pixel Dungeon by Evan Debenham. Features 50 dungeon floors with bosses every 5 floors, and full autonomous real-time gameplay.
Poppins
Child-development digital therapy platform — a client project.
A telehealth platform delivering remote digital therapy and assessments for child development. Built end-to-end at Weaversoft.io for Femi.
GainGuard.ai
Backend & cloud infrastructure for an AI sports-safety wearables platform.
Designed the FastAPI microservices backend with a Django API gateway, built GCP cloud infrastructure and load balancing, integrated the React frontend, and performed ISO 27001-aligned security hardening.
Document RAG
Local document RAG with hybrid retrieval and real inline citations.
Self-hosted RAG over your own documents, audio, video, and web pages. Hybrid dense + sparse search (BGE-M3) with cross-encoder reranking, multilingual EN/RU/HE, and streamed answers with clickable citations that open the source chunk.
About Me
Senior full-stack & AI engineer with 5+ years shipping production applications end-to-end — from React frontends through Django/FastAPI backends to Kubernetes deployment. I focus on AI/LLM-powered systems: RAG pipelines, document-processing, and self-hosted GPU inference.
I own system architecture end-to-end, mentor developers, and lead production troubleshooting. I enjoy building systems that are clean, maintainable, and performant — trained as a Practical Engineer in Electronics (ORT College).
Fun fact: Linus Torvalds is my godfather.
Experience
Senior Full-Stack & AI Engineer
@ Weaversoft.io
- Built a real-time medical platform with React, Django, and MySQL supporting 50–500 concurrent clinical sessions
- Built a document-processing ML system combining OCR, fine-tuned LLMs, and algorithmic extraction
- Developed and deployed an early RAG-based tool-calling system for chat agents (pre-LangChain/Flowise) on GPT-3: similarity search over an indexed CSV of tool names, descriptions, and Python code, with the matched row's code extracted and executed via a PostgreSQL vector store
- Developed a healthcare conversational RAG agent with dynamic retrieval and domain knowledge integration
- Architected large-scale multi-tenant SaaS platform in Django orchestrating Docker microservices on k8s
- Managed a team of 6 developers and 2 DevOps engineers across the full SDLC
- Developed a self-hosted k3s platform for managing and deploying custom applications with Helm, providing a UI for system resource management and app deployment with SSO, LDAP, and an Identity provider. Managed SDLC environments for agents on the platform with Gitea, code-server, and opencode, powered by self-hosted models on vllm/llama.cpp containers.
- Developed secure company agents using MCPs integrating with various aspects of the company, featuring user/role management under admin user, enabling documentation (company data lake), task management, and internal/client meeting summarizations with context recalling.
- Obtained ISO27001 certification for Weaversoft.io