hiring for backend · /be
Hiring a Backend Engineer?
Event-driven services, multi-tenant Postgres, and realtime systems in Go and Python.
I design backends for isolation and change: event-driven microservices, schema-per-tenant Postgres with RLS, Go realtime fan-out with sub-millisecond broadcast, and sub-100ms pricing engines. Everything observable with OpenTelemetry, and deployable independently.
Backend EngineerSoftware Engineer (Backend)Distributed Systems EngineerAPI / Platform Engineer
<100ms
personalised surge-pricing engine
sub-ms
circle-scoped WebSocket broadcast
4-tier
SLA escalation state machine
RLS
schema-per-tenant isolation
// 01 — relevant experience
Where I did this before.
- Own infra provisioning across dev/staging/prod with Terraform (VPC, EC2, RDS, IAM) as Infrastructure-as-Code; GitHub Actions pipelines enforce plan-before-apply, drift detection, and SSM/OIDC-backed secrets with zero long-lived credentials.
- Architected infra-ops, a centralized multi-environment IaC platform: S3 remote state, DynamoDB state locking, reusable Terraform modules (networking/compute/storage/IAM) and Ansible roles for PostgreSQL/PostGIS, Nginx+Certbot and PM2 deploys.
- Built lgtm-sidecar, a resource-capped observability stack (Grafana Alloy, Loki, Tempo, Prometheus, Pyroscope, Alertmanager) as a decoupled EC2 sidecar with distributed tracing and rate-capped tail sampling.
- Built an AI-powered BaaS platform generating full-stack e-commerce apps (RESTful APIs, DB schemas, frontend) from prompts, deployed on AWS (EC2, ECS, Lambda, Step Functions, DynamoDB, Valkey, CloudWatch).
- Re-architected monolithic AWS infra into event-driven, containerized microservices, cutting monthly cloud spend by 76% with a zero-downtime migration.
- Owned observability and on-call readiness with CloudWatch alarms and structured logging, reducing mean-time-to-detect and accelerating root-cause resolution.
- Built a sub-100ms dynamic surge-pricing engine that personalizes user rates from real-time availability, spend history and geolocation.
- Refactored a monolith into event-driven microservices with async communication, improving isolation, independent deployability and integration with dependent systems.
- Collaborated cross-timezone with a distributed US team in an agile workflow: sprint planning, code review and incremental rollout of API and pricing changes.
2025-11 → present
Software Engineer
RaghuvarTech (Raghuvar IT Solutions LLP)
Own infrastructure, CI/CD and observability for a multi-service platform; lead architecture reviews across backend, infra and agentic-AI systems.
2025-06 → 2025-10
Software Development Engineer
InkHub.in / BRMH.ai
Built an AI-powered backend-as-a-service that generates full-stack e-commerce apps from prompts, and re-architected its AWS footprint.
2024-11 → 2025-05
Software Development Engineer (Remote)
EatTogether & Co.
Backend engineer on a US food-tech team: real-time pricing, monolith-to-microservices refactor, cross-timezone agile delivery.
// 02 — selected case studies
Proof, with numbers.
// 03 — skills for this role
The toolbox.
Languages & frameworks
- Python / FastAPI
- Go
- Node.js / Express
- TypeScript
- SQL
- C++
Architecture
- Event-driven microservices
- multi-tenant design
- REST / gRPC / GraphQL
- WebSockets / SSE
- idempotency & retries
- state machines
Data & messaging
- PostgreSQL (RLS, locking, PostGIS)
- Redis
- Kafka
- DynamoDB
- Milvus
- caching strategies
Operations
- OpenTelemetry
- LGTM stack
- Docker / Kubernetes
- CI/CD
- load & latency testing
// 04 — relevant writing
How I think about it.
012 min
Advanced RAG Architecture: Graph, Vector, Redis, and PostgreSQL
Combining Graph databases, Vector databases with Reciprocal Rank Fusion (RRF), Redis caching, and PostgreSQL to build a state-of-the-art RAG system.
ragaidatabasesvector db
022 min
Designing Multi-Tenant Database Architectures
A comprehensive guide to the different approaches for building multi-tenant databases in SaaS applications.
databasearchitecturesaasmulti-tenancy
033 min
Sub-millisecond WebSocket fan-out in Go: the Vedico realtime service
How a single Go service pushes GPS telemetry from hundreds of field technicians to NOC dashboards in under a millisecond, with back-pressure, per-region isolation and no shared locks on the hot path.
gowebsocketsrealtimebackend
// 05 — recruiter faq
The questions you were going to ask.
What's the most demanding backend you've built?
Vedico's realtime layer: a Go service broadcasting GPS telemetry per 'circle' with sync.RWMutex-guarded connection maps and channel multiplexing, plus geofencing, pessimistic locking for anomaly detection and a 4-tier SLA escalation state machine.
How do you approach multi-tenancy?
Schema-per-tenant with Postgres RLS and a dynamic search_path, unified RBAC on top, and per-tenant observability via span baggage so noisy tenants are visible without high-cardinality labels.
How do you make services independently deployable?
Async communication over events, explicit contracts, idempotent consumers, health-check-gated rollouts and per-service pipelines; I did this refactor twice (EatTogether and InkHub).
Where are you based and can you work remotely?
Based in Mumbai, India (IST). Remote-first; I've shipped with a Seattle team across time zones and I keep 4+ hours of overlap with US Pacific or any EU zone.
What is your notice period?
Currently employed; I can start within 30 days, sooner for contract work.
Full-time or contract?
Open to full-time remote roles and to contract/consulting engagements. Relocation is possible for the right role.
// talk to me about a backend role
Book a 15-min call
Pick a time that fits you, or just
email me at hi@nikhilsahu.dev