Open to software & AI engineering roles
I build distributed backends, LLM systems, and automation pipelines designed for reliability in production.
Biniyam Seid — Software & AI Engineer. 5+ years across distributed systems, LLM applications, and cloud infrastructure. I care about correctness, observability, and the trade-offs that survive contact with real traffic.
Also taking selected freelance engagements via Upwork and Freelancer.
- yrs engineering
- 5+
- Postgres / AWS p99 gain
- 25%
- ops latency automated away
- 40%
- engineers mentored (A2SV)
- 30+
Selected work
Case studies, not screenshots
Each project below ships in production or near-production with measurable impact. Click through for architecture, decisions, and results.
How I work
Engineering principles I optimize for
These are the defaults I bring into every codebase — whether the engagement is a full-time team, a research project, or a freelance build.
Type-safe by default
TypeScript strict mode, validated schemas at every boundary, no `any` in the seams. The compiler is the first reviewer.
Boring infra over novelty
PostgreSQL, well-supported queues, and small Dockerfiles before bespoke distributed magic. Pick the tool that is still working at 2am.
Observable from day one
Structured logs, request IDs, latency / cost dashboards, and traces wired before the first deploy — not bolted on after the first incident.
Tests at the seams
Integration tests at module boundaries and contract tests for external APIs. Unit tests where pure logic earns them.
Measure before optimizing
Real query plans, real p99s, real flamegraphs. Most “performance” problems are actually data-shape or schema problems.
Small, reversible changes
Atomic commits, feature flags, and migrations that can roll back. Production should never be one wrong button away from a long night.
Stack
Tools I reach for in production
The list is long; the rule is simple — pick the boring, well-supported tool unless the problem demands otherwise.
Languages
Frameworks
ERP / Automation
AI / ML
Cloud / DevOps
Data
Focus areas
What I build, in production
The technical surface I work across. Outcomes are linked to specific case studies — each is real work, not a stack list.
AI / LLM systems
FastAPI + LangChain / LangGraph agents, RAG pipelines, evaluation harnesses, embeddings on AWS. Shipped SMSAI and research engineering at SingularityNET.
→ Production LLM apps on AWS
Distributed backends
Next.js, Go, Node, FastAPI. Service decomposition, queues, caches, schema design. Kubernetes, Docker, GitHub Actions, infra-as-code with Terraform.
→ Full-stack systems end-to-end
Databases & performance
PostgreSQL indexing and query tuning, DynamoDB modeling, p99 latency work, and the boring observability that catches regressions early.
→ 25% Postgres / AWS p99 gain
Workflow automation
n8n / Zapier pipelines integrating CRM, comms and ERP. Twilio SMS, voice and WhatsApp inside business workflows. Used heavily in client engagements.
→ 40% manual ops removed
Let's talk
Hiring, collaborating, or building something hard?
I'm open to senior software and AI engineering roles, and I also take on a small number of freelance engagements at a time. Either way — tell me what the problem is and what good looks like.