Tusdi AI
Full-Stack Engineer · Next.js, FastAPI, PostgreSQL, Python
Greenfield medical AI platform. The system parsed medical documents, tracked conditions, symptoms, and visit histories so patients and doctors could have more productive interactions. I worked on both the frontend and backend, and spent a lot of time navigating the challenges of AI indeterminism in a domain where accuracy expectations were very high.
Codevalet
Full-Stack Engineer · Next.js, FastAPI, Redis, Celery, Kubernetes, Helm
Code intelligence platform for understanding and documenting repositories. The existing stack needed reworking, so I took initiative to migrate the UI and backend to Next.js and FastAPI. Also helped rebuild the AI pipeline with Celery and Redis Streams, and put together Helm charts for multi-cloud Kubernetes deployment.
XLRT
Freelance Full-Stack Engineer · Next.js, Python, RabbitMQ, PostgreSQL
Finance platform using LLMs to analyze company reports and help banks make credit approval decisions. Worked on dynamic dashboards for company-level financial data with server-driven layouts, and integrated chat and report generation via RabbitMQ.
ICOMS
Freelance Full-Stack Engineer · Next.js, Laravel, FastAPI, Kubernetes
Visual website builder in the space of Webflow and Framer. Contributed a component library for the visual editor and worked on automated custom domain assignment through Kubernetes — SSL certificates, ingress rules, and routing. Also did early experiments with GPT-3.5 for UI generation.
YoFi (BotNot.io)
Lead Backend & DevOps · Python, Node.js, AWS, GCP, MongoDB
Anti-fraud AI platform for e-commerce. I was part of the backend and DevOps side — helping manage 20+ microservices, contributing to ML data pipelines with Airflow and BigQuery, and working on serverless applications across AWS and GCP. Went through the challenges of automated deployment with SAM templates, CDK, and SST. The team maintained 99.9% uptime.
scienceweb.uz
Lead Full-Stack Engineer · Vue.js/Nuxt, Laravel, Python, Node.js
Research publishing platform for Uzbekistan. I built most of the stack — frontend, backend, and integrations with Scopus, Web of Science, ORCID, Google Scholar, and Crossref for automatic DOI resolution. Originally wrote it in Vue 2.6, later rewrote it with Vue 3 and Composition API. The platform grew to 14,000+ publications from 4,000+ researchers.