Blessing Softtech
Website:
blessingonline.com
Company:
https://www.linkedin.com/company/blessingsofttech
Industries: IT Services and IT Consulting
Job details:
Staff Backend Engineer — Real-Time Voice InfrastructureFonazo (Blessing Softtech) · Pune, India · Full-time, on-site/hybrid
What you will own- Concurrency, isolation, and latency under load. Many simultaneous live voice sessions, with predictable p95 turn latency as volume rises. This is the hardest problem in the product and your primary mandate.
- Production operability. Observability stack from zero (metrics, tracing, alerting, on-call), incident response, load management, backend CI/CD, backups/DR.
- Money-path reliability. Payment flows (PhonePe/Stripe), the credits ledger, WhatsApp Business messaging, and SMS — the customer-facing paths that must never silently fail.
- Cost-per-minute engineering. Every call burns STT/LLM/TTS/telephony vendor spend. You'll instrument and drive unit economics down without degrading latency.
Hard requirements — every one is verifiable, and we will verifyApply only if all of the following are true. We check these in the deep-dive interview and via reference calls with people who worked the same systems.
- You have operated (not just built) a production system that sustained ≥500 concurrent long-lived connections (WebSocket, SIP/RTP, streaming gRPC, or similar) — you carried the pager for it.
- You can state, from memory, the p50/p95/p99 latency numbers of a system you owned, what the bottleneck was, and what you changed to move them.
- You have been the incident commander or primary responder for at least one Sev-1 in a system serving real paying users, and can walk through the timeline, the fix, and the postmortem action items.
- You have written or materially owned payment/ledger code where a bug meant real money lost — idempotency, reconciliation, double-spend prevention.
- Deep, hands-on Python or Go backend experience, Postgres at scale (connection pooling, locking, migration discipline), Kubernetes in production, and IaC.
- 6+ years backend, with at least 2 at a company with a large active user base and high daily traffic — you've already met the class of problems that only appear at volume.
Strong signal: real-time voice/video (WebRTC, SIP, media servers), telephony (Twilio/Exotel/FreeSWITCH), streaming ML inference pipelines, FinTech infra in India (UPI/PhonePe ecosystem).
What we are explicitly NOT looking for- Strong 0→1 stories with no lived experience of what breaks at volume. We have 0→1 covered. We need 1→N.
- "Led a team that…" without personal, hands-on ownership. Staff here means you write code and carry the pager.
- Framework-collector résumés. We care what broke at 3 a.m. and what you did about it.
Click on Apply to know more.