Skip to content
● akshay.nanavare
menu

happy to talk backend, AI & system design

Akshay Nanavare

I led our observability platform's ClickHouse migration to Altinity: 50+ live tenants, zero downtime.

Senior engineer at Last9 working on observability: logs, traces, APM, ClickHouse. Before that, payments at Tazapay.

live tenants migrated

50+

ClickHouse Cloud → Altinity, zero downtime

events / min peak

960M

IPL & India vs Pakistan peaks

report generation

15m → <1m

Go + CDC pipeline at Tazapay

in production

7 yrs

shipping backends since 2019

featured

Last9 · Zero-downtime ClickHouse migration

PROBLEM
Compute cost grew with every tenant we added, and customers rely on ingestion, search and alerts around the clock, so downtime was not an option.
MY ROLE
I designed the migration strategy, mentored by our CTO, and three engineers joined for the rollout.

trace: akshay.career

Full experience →
  1. netscout 2019–21
  2. parallel-wireless 2021–22
  3. tazapay 2022–24
  4. last9 2024–now

selected work

All projects →

shell: ask my resume anything

akshay@portfolio:~
Welcome. Type 'help' to see what's here, or tap a suggestion below.

Experience

  1. Aug 2024 – Present

    Senior Software Engineer · Last9

    Remote

    • Led the zero-downtime migration of 50+ live tenants from ClickHouse Cloud to Altinity, significantly cutting compute cost
    • Kept the ingestion pipeline (OpenTelemetry Collector → Kafka → ClickHouse) stable at up to 960M events/min (logs and traces) during IPL and India vs Pakistan traffic peaks
    • Designed and built the APM backend, one of the product's core features, from MVP onwards
    • Built ingestion-time extract, remap and drop rules, plus user-defined physical indexes that route filtered data (say, staging or production logs) into its own table
    • Built cold storage that backs up logs to the customer's own S3 or GCS bucket and rehydrates them on demand, so archived logs are queryable again
    • Automated tenant onboarding, so a SaaS signup or BYOC (bring your own cloud) setup provisions the tenant's infrastructure on Kubernetes
    • Led core Logs and Traces features from design to production
    • Optimized ClickHouse query performance for large-scale logs and traces
    • Built a ClickHouse SQL query generator supporting filters, parsers (regex, JSON, logfmt), and transformations
    • Architected an event-driven S3 ingestion pipeline for logs (near real-time)
    • Built an MCP server for real-time debugging from Cursor, VS Code, and Claude
  2. Jul 2022 – Aug 2024

    Backend Engineer · Tazapay

    Remote

    • Lead contributor to Account, Report, KYB, Notification, Event, and OpsDashboard microservices (Go, gRPC, 92% unit test coverage)
    • Cut transaction report generation from 15 minutes to under 1 minute with a new Go service + CDC (Kinesis) architecture
    • Transaction monitoring: 40% faster processing, 15% less memory, 10k–30k transactions/day
    • Designed the PayFac model for the Rapyd provider; refactored card routing in the Router service
    • Built an ops control panel in Go for managing merchant payment methods; reduced ops tickets by 80%
    • Built merchant RBAC and custom notifications for async payment methods (bank/wire transfers)
  3. Jun 2021 – Jul 2022

    Software Engineer · Parallel Wireless

    Pune

    • Built microservices in Go with REST APIs; enabled multi-client support in a core service
    • Owned the Alarms/Events microservice; supported CI/CD automation
  4. Aug 2019 – Jun 2021

    Software Engineer · Netscout

    Pune

    • Built an automation framework that generates dynamic web reports
    • Built HTTP APIs for standalone C++ tools
    • Restructured makefiles for parallel compilation, speeding up builds

Selected work

  • Zero-downtime ClickHouse migration (Last9): 50+ live tenants migrated with zero downtime and significantly lower compute cost.
  • APM backend (Last9): Became one of the product's core features; the MVP contributed to product-market fit.
  • Ingestion control plane and physical indexes (Last9): Customers trim data before paying to store it, and indexed queries scan only the data they need.
  • Cold storage and rehydration (Last9): Long retention at object-storage prices, with archived logs queryable again whenever they're needed.
  • Tenant onboarding automation (Last9): New tenants, SaaS or BYOC, come up without manual infra work.
  • Event-driven S3 log ingestion (Last9): Logs from S3 become queryable in near real time.
  • ClickHouse SQL query generator (Last9): Powers log querying in the product, with query performance tuned for large-scale logs and traces.
  • MCP server for production debugging (Last9): Real-time debugging from Cursor, VS Code, and Claude.
  • Transaction report pipeline (CDC) (Tazapay): Report generation went from 15 minutes to under 1 minute.
  • Ops control panel (Tazapay): 80% fewer ops tickets.
All case studies →

Skills

Languages
  • Go
  • Python
  • SQL
  • C++
Databases
  • ClickHouse
  • PostgreSQL
  • MongoDB
  • DynamoDB
Infra
  • Docker
  • Kubernetes
  • Terraform
  • Redpanda (Kafka)
  • Git
Cloud & other
  • AWS
  • GCP
  • OpenStack
  • gRPC
  • MCP servers
  • DSA

Education

BE Computer Science, Govt. College of Engineering and Research, Avasari, Pune (2015 – 2019)

↑↓ navigate · enter select · esc close