DevOps Engineer (Senior / Platform Lead)

About Company SuperTrading Research Capital (STRC) (www.supertrading.in

) is a bootstrapped, profitable quantitative trading and technology firm headquartered in Mohali, India, deploying proprietary capital across equities, derivatives (multi-leg options, futures), commodities, and currencies in Indian capital markets. Founded by an IIT alumnus and serial capital-markets entrepreneur, SuperTrading is built from the ground up by engineers and quant researchers from India's top technical institutions. We are not a retail brokerage or a consumer fintech app with a trading skin—we are a research-driven, latency-obsessed proprietary firm that designs, backtests, and deploys high-frequency and algorithmic strategies, operating our entire high-throughput execution engines, tick-level market data pipelines, co-located matching servers, and real-time risk systems completely in-house. Our code trades live capital every single market day, where system performance, execution speed, and architectural reliability are measured directly in rupees and market alpha rather than vanity engagement metrics.

You will lead the operational evolution across infrastructure, data pipelines, media processing, and deployment automation. Specifically, you will:

• Kubernetes Orchestration & Workload Migration: Migrate the ~25 Go services from per-host Docker Compose configurations to Kubernetes (EKS or self-managed) using Helm charts, executing the cutover with zero downtime during active trading hours. Make deliberate architectural determinations on which services migrate to Kubernetes and which remain on dedicated bare VMs for latency, hardware pinning, or isolation reasons.
• High-Throughput Kafka Operations: Own Apache Kafka in production: execute cluster sizing, broker provisioning, partition planning for tick-level throughput, ISR (in-sync replica) management, retention policies, consumer-group lag alerting, and safe zero-downtime broker upgrades.
• In-Memory Caching & Redis Architecture: Own Redis 7/8 instances across both platforms: configure persistence (AOF), memory limits (noeviction), failover clustering (Redis Sentinel or Cluster), and actively monitor BullMQ queue performance and worker throughput.
• PostgreSQL & TimescaleDB Reliability: Operate PostgreSQL and TimescaleDB: design and execute automated backup strategies and live restore drills, manage read replication, monitor and optimize cross-region replication latency between Mumbai (ap-south-1) and Hyderabad (ap-south-2), and tune connection pooling (e.g. PgBouncer).
• Real-Time Media & LiveKit Scaling: Self-host, configure, and scale LiveKit for WebRTC rooms: manage TURN/STUN servers, UDP port configurations, dynamic SFU node autoscaling, egress recording pipelines, and HLS mirror streaming distribution via CloudFront.
• AI / LLM & Vector Search Infrastructure: Productionize AI infrastructure: manage ChromaDB as an enterprise-grade service (automated backups, persistence, authentication, scaling beyond single nodes) or lead migration to pgvector/Qdrant; keep Vertex AI Vector Search indexes performant and cost-optimized; manage OpenRouter gateway keys, per-service spend limits, rate-limit backoff (429 handling), model fallbacks, and comprehensive token, latency, and cost telemetry dashboards.
• Unified Observability & Market-Hours Alerting: Build a single consolidated observability platform: configure Prometheus to scrape every microservice, Kafka, Redis, Postgres, and LiveKit exporter; build high-utility Grafana dashboards; configure Alertmanager rules with PagerDuty/on-call routing; and deploy centralised logging (Loki or equivalent) to replace disparate per-container log volumes.
• Continuous Delivery & SSH-Free Pipelines: Evolve existing GitHub Actions CI pipelines into automated Continuous Delivery: engineer multi-stage arm64/amd64 image builds, automated environment promotion stages, zero-downtime rollbacks, and eliminate manual SSH server deployments.
• Infrastructure as Code & Secrets Management: Codify all cloud infrastructure using Terraform and eliminate insecure plaintext .env files by migrating to AWS Secrets Manager or HashiCorp Vault.
• Edge Security & Network Hardening: Harden edge perimeters: automate Caddy and Ingress TLS certificates, enforce private-only ports, configure IPv6 egress proxies, audit security groups, and enforce strict IAM least-privilege policies.
• Cloud FinOps & Cost Optimization: Establish unified cost visibility across AWS, GCP, and OpenRouter, eliminating zombie infrastructure and right-sizing underutilized compute and database tiers.
• Operational Rigor, Runbooks & Post-Mortems: Author comprehensive operational runbooks, take primary ownership of incident response during Indian trading hours (09:15–15:30 IST), and lead blameless post-mortem investigations following any production disruption.
• Must-Have Technical Skills & Evaluation Criteria

Candidates must possess 5–7 years of hands-on experience in DevOps, SRE, or Platform Engineering with proven production ownership of each core domain. Note: During technical interviews, we will evaluate specific production incidents, architectural trade-offs, and debugging decisions rather than theoretical trivia or certifications.

Technical Domain What We Expect at the 5–7 Year Senior Level Kubernetes 2+ years running production clusters (EKS or self-managed). Deep experience with Helm charts, HPA, PodDisruptionBudgets, network policies, arm64 node pools, and stateful workloads. Proven track record of migrating legacy services into Kubernetes.

Docker & Packaging Multi-stage Docker builds for Go and Node.js, arm64 vs. amd64 multi-arch cross-compilation, container image size minimization, vulnerability (CVE) scanning, and Docker Compose for local/single-host environments.

Apache Kafka Hands-on production cluster operations: broker sizing, partitioning topology for high throughput, replication factors and ISR, consumer-group lag alerting, MirrorMaker/replication, and zero-downtime rolling upgrades. Practical understanding of idempotent producers and exactly-once processing trade-offs.

Redis Architecture Production administration of persistence mechanisms (RDB vs. AOF), memory limits (maxmemory eviction policies), high-availability failover (Redis Sentinel or Redis Cluster), Pub/Sub vs. Streams vs. BullMQ queues, and debugging memory latency under heavy load.

WebRTC & Media Servers Production deployment and operation of real-time media servers (LiveKit, mediasoup, Janus, or equivalent). Deep familiarity with TURN/STUN traversal, UDP/firewall configuration, SFU scaling, egress recording pipelines, and HLS/RTMP packaging.

OpenRouter / LLM Gateways Production traffic management via OpenRouter or LLM gateways (LiteLLM, Portkey): multi-service API key provisioning, per-team spend caps, rate limit (HTTP 429) backoff, dynamic model fallbacks, sanitized request logging (zero prompt leakage), and granular cost attribution per feature.

Vector Databases Production operational experience with at least one vector database (ChromaDB, pgvector, Qdrant, Weaviate, Milvus, or Vertex AI Vector Search): persistence, automated backups, index rebuilding, embedding pipeline workers, memory dimension sizing, query latency profiling under load, and evaluating self-hosted vs. managed trade-offs.

Prometheus & Metrics Dynamic service discovery, exporter deployment (node_exporter, kafka_exporter, redis_exporter, postgres_exporter), recording rules, metric cardinality governance, Alertmanager routing, and familiarity with long-term storage solutions (Thanos / Mimir a strong plus).

Grafana Dashboards Designing actionable operational dashboards: per-service RED (Rate, Errors, Duration) and USE (Utilization, Saturation, Errors) frameworks, SLO/SLA tracking panels, alert annotation overlays, and dashboard provisioning via code/git.

Additional Core Engineering Requirements

• AWS Cloud Architecture: Deep hands-on experience across EC2 (including Graviton instances), VPC networking, IAM security, S3 storage, CloudFront CDN, RDS or self-managed PostgreSQL, and multi-region connectivity between Mumbai and Hyderabad.
• Linux Systems & Performance Debugging: Expertise with Ubuntu server environments, systemd service management, kernel and network parameter tuning, and low-level performance debugging under high load.
• CI/CD with GitHub Actions: Configuring secure workflows, caching strategies, and operating self-hosted runner infrastructure.
• Infrastructure as Code: Writing modular, maintainable, state-managed Terraform code.
• Polyglot Scripting & Code Comprehension: Fluency in Bash and either Go or Python; ability to read Go and TypeScript codebase logic comfortably enough to debug application-level issues alongside developers.
• Operational Ownership & Market Hours On-Call: Calm, disciplined approach to on-call duties for mission-critical trading infrastructure where real money is on the line.
• Interview Process & Evaluation Stages

Our hiring process is straightforward, respectful of your time, and focused directly on real-world engineering challenges:

• Stage 1: Introductory Discussion (30 mins): Initial conversation with the founding engineering team to discuss your background, expectations, and alignment with the role.
• Stage 2: Systems & Production Deep-Dive (60 mins): In-depth review of real-world production incidents you resolved, architecture decisions made around Kubernetes, Kafka, Redis, or WebRTC, and practical troubleshooting methodologies.
• Stage 3: Hands-On Architecture / Debugging Exercise (60 mins): A practical, pragmatic scenario focusing on service containerization, pipeline optimization, or diagnosing infrastructure bottlenecks.
• Stage 4: Culture & Team Alignment (30 mins): Discussion on working directly with core developers, market-hours on-call expectations, compensation, and offer rollout.

Skills: docker,kubernetes,microservices,aws,postgresql,redis,node.js,typescript,kafka