sachin kant
Say hello

Backend & infra engineer · Superleap · Bengaluru

Backend & infrastructure engineer.

I build distributed systems in Go, run them on Kubernetes, and wire AI agents into them.

Trace
sk-404
Spans
13
Status
running
Now
2 OCT 2026
TRACE · GET /sachin-kantrunning · 13 spans
  • career sachin-kant
    running
  • dypcoe-pune b.e. cs
    2021 – 2025
  • side-project loadbalancer
    go · rr + wrr
  • superleap software-engineer
    2024 → now
  • jobs river
    go · postgres
  • platform aks + ci/cd
    deploys · uptime
  • platform temporal-cluster
    self-hosted · k8s
  • workflows temporal
    go · temporal
  • ai-tools mcp-server
    multi-tenant
  • auth oauth 2.0 + rbac
    per-tool rbac
  • audit kafka log
    every call
  • ai call-intelligence
    llm · rag
  • rag retrieval
    transcripts
click a span to open it
NOW · 2 OCT 2026
  • Education
  • Superleap
  • Side project
Prefer reading? Scroll ↓

Work

In production

The systems that page me, in no particular order.

Org
Superleap
Role
Backend & Infra Engineer
Since
Jun 2024
Location
Bengaluru
01InfrastructureRuns production: AKS, CI/CD and observability.

Manage production on Azure Kubernetes Service: deploys, scaling, observability and reliability for the core services.

Moved deploys from manual VMs to Kubernetes, with multi-environment Jenkins pipelines that ship many times a day.

Autoscaling policies plus monitoring with Grafana, Prometheus and SigNoz.

Depends onAKS · Kubernetes · Jenkins · Prometheus · Grafana · SigNoz

02Durable workflowsTemporal workflows, and the cluster they run on.

Built Temporal workflows in Go: retries, timeouts and long waits live in the workflow, not in cron jobs.

Run the self-hosted Temporal cluster they execute on, which carries the product’s async and long-running work.

Depends onTemporal · Go · Kubernetes

03Async jobsBackground job queues on River.

High-performance Go APIs and background job queues on RiverQueue, tuned for throughput.

Depends onGo · River · PostgreSQL

04AI toolingMCP server: tenant routing and a tool registry.

Built an MCP server for the Superleap CRM with multi-tenant routing, custom MCP creation and an extensible tool registry.

Tenants add their own custom tools that call external services.

OAuth 2.0 authentication and RBAC-based authorisation for every tool.

Kafka-backed audit logging for every tool execution.

Depends onGo · MCP · OAuth 2.0 · RBAC · Kafka

05Call intelligenceLLM + RAG insights from sales calls.

Multi-tenant call intelligence with dynamic prompt orchestration for tenant-specific insights.

RAG over call transcripts, so insights and reports are grounded in what was actually said on each tenant’s calls.

Transcription, sales call analytics and pitch reports, plus compliance-violation flags.

Depends onGo · LLMs · RAG · Multi-tenant

Projects

Side processes

Low priority, never killed.

loadbalancer

Public

Go · 2025 · Round-robin. Weighted. Self-healing.

A load balancer in Go with a pluggable core: new balancing strategies and health checks slot in without touching it.

  • + rr / weighted rr
  • + health checks
  • + new algorithm = new file
github.com/sachin-404 →
client → lb:8080
backend-1w=3
backend-2w=1
backend-3skipped

Stack

The stack, top to bottom

Grouped the way a request travels through it.

Languages

  • Go
  • Python
  • SQL

Interfaces

  • REST APIs
  • MCP
  • OAuth 2.0

Orchestration

  • Temporal
  • River
  • Kafka

Data

  • PostgreSQL
  • Redis

Platform

  • Kubernetes · AKS
  • Docker
  • Terraform
  • ArgoCD
  • Jenkins

Cloud

  • Azure
  • AWS
  • Google Cloud

Observability

  • Prometheus
  • Grafana
  • OpenTelemetry
  • SigNoz

Contact

Say hello

No ticket required. Email is the fastest way to reach me.

sachinkant404@gmail.com
GitHub ↗LinkedIn ↗