buckwheat8848
Remote: Yes
Willing to relocate: No
Technologies: TypeScript, Node.js, NestJS, Python, FastAPI, PostgreSQL, Redis, Next.js, Flutter, Docker, Kubernetes on GCP, Cloudflare, GitHub Actions, LLM APIs, pgvector
Résumé/CV: on request, code at https://github.com/kafle1
Email: [email protected]
Full stack engineer, 5 years in production, most of it backend. I own a NestJS and PostgreSQL backend for a US company and run its deployments on Kubernetes on GCP.
I also built and run a business operations platform alone: 490 API routes on a 150-table schema. It deploys itself every 2 minutes and rolls back when a health check fails.
2 years of LLM features in production, voice agents and retrieval among them.
Side project: https://github.com/kafle1/relay runs traffic signals from YOLO vehicle counts. In simulation it cut delay by 23 to 56% against a fixed plan.
I work 14:00 to 23:00 Nepal time, which is the UK and EU day and US Eastern mornings. Contract or full-time. I invoice as an independent contractor.
Remote: Yes, worldwide. I cover the full UK and EU working day, and US East mornings.
Willing to relocate: No
Technologies: TypeScript/Node (NestJS, Next.js, SvelteKit, React), Python (FastAPI), PostgreSQL, Redis, Docker, Kubernetes on GCP, GitHub Actions, AWS, Azure, Cloudflare. LLM work in production for the last two years: retrieval over pgvector, multi-agent workflows, eval harnesses, voice AI.
Résumé/CV: on request by email. Code: https://github.com/kafle1
Email: [email protected]
Five years full stack, strongest on the backend. Right now I am the engineer who builds and runs the services at a small US company: NestJS APIs on Postgres, the Next.js and SvelteKit front ends on top, and the delivery side too (Docker images, k8s on GCP, Cloudflare, CI/CD, rollbacks). The last two years of that included LLM features people actually use, not demos: retrieval pipelines, agent workflows, a voice sales bot, and the eval harness that keeps a non-deterministic pipeline honest with a golden set and step traces.
Looking for contract or full-time, and I prefer work scoped to deliverables over seat time. I invoice from Nepal as an independent contractor. If you want to read code before replying: a queue-weighted max-pressure traffic signal controller validated on real junction footage (github.com/kafle1/relay), and a single-camera cricket LBW review tool (github.com/kafle1/pocket-drs).
Remote: Yes, worldwide. 4-5 hours of overlap with Europe, US East mornings.
Willing to relocate: No
Technologies: TypeScript/Node (NestJS, Next.js, SvelteKit, React), Python (FastAPI), PostgreSQL, Redis, Docker, Kubernetes on GCP, GitHub Actions, AWS, Azure, Cloudflare. LLM work in production for the last two years: retrieval over pgvector, multi-agent workflows, eval harnesses, voice AI.
Résumé/CV: on request by email. Code: https://github.com/kafle1
Email: [email protected]
Five years full stack, strongest on the backend. Right now I am the engineer who builds and runs the services at a small US company: NestJS APIs on Postgres, the Next.js and SvelteKit front ends on top, and the delivery side too (Docker images, k8s on GCP, Cloudflare, CI/CD, rollbacks). The last two years of that included LLM features people actually use, not demos: retrieval pipelines, agent workflows, a voice sales bot, and the eval harness that keeps a non-deterministic pipeline honest with a golden set and step traces.
Looking for contract or full-time, and I prefer work scoped to deliverables over seat time. I invoice from Nepal as an independent contractor. If you want to read code before replying: a queue-weighted max-pressure traffic signal controller validated on real junction footage (github.com/kafle1/relay), and a single-camera cricket LBW review tool (github.com/kafle1/pocket-drs).