Bharat Verma

Pune, India · Working across IST and US hours

Bharat Verma

I build the platform organisations that financial-grade systems run on.

Open to Director / Head of Engineering rolesIndia · Singapore · UAE

Sixteen years across payments, enterprise SaaS and cloud infrastructure. I've taken PHP monoliths into Go and Java services, stood up real-time data platforms, and run card payment systems where a two-second regression is a business incident rather than a graph.

Most of my work sits where architecture meets organisation design: deciding what a platform team owns, what it exposes to other teams, and how it stays reliable when the company grows faster than the codebase.

Currently
Director, Platform Engineering — AssetMark
Previously
Razorpay · Guidepoint · Acquia · MobiKwik · Naukri.com
16 yrs
Engineering, 10 of them leading teams
15K TPS
Peak throughput on payment systems owned
$60–90M
Daily transaction volume under ownership
40+
Engineers hired across levels since 2016
  • Payments & fintech platforms
  • Distributed & event-driven systems
  • Cloud-native platform engineering
  • GCC build-out & site scaling
  • SRE, observability & FinOps

Profile

What I actually do

Six areas cover most of my work. The through-line is the same in each: make the expensive decisions early, keep the boring things reliable, and give engineers a system they can reason about.

I stay close to the code — architecture reviews, design docs, the occasional production trace — because leadership decisions get worse the further you drift from the thing you're deciding about.

01

Platform engineering

Shared services, API-first contracts and event-driven backbones that several product teams depend on. The platform is a product; developer experience and extensibility count as much as the feature that shipped this quarter.

02

Payments and regulated systems

Card, recurring and e-mandate flows under PCI-DSS and RBI PPI constraints. Idempotency, reconciliation, partner failover and the operational discipline that keeps a payment platform honest at 15K TPS.

03

Cloud-native and FinOps

Kubernetes, Terraform and CI/CD on AWS and Azure, with cost treated as a first-class metric — including $660K of annual COGS removed at Acquia and per-transaction logging cost cut by two-thirds at Razorpay.

04

Real-time data

Kafka, Spark Streaming and Apache Pinot for operational analytics, plus Databricks pipelines for discovery products. Built for people who need to answer a question before the incident bridge fills up.

05

Org design and GCC build-out

Multi-pod organisations of 20–30 engineers, hiring across levels, and turning an offshore office into a Global Capability Center with genuine ownership rather than ticket execution.

06

AI-assisted delivery

Claude Code, Cursor and n8n adopted as team practice rather than a pilot — code review, refactoring, documentation and test generation — alongside LLM features shipped into production.

How the loop runs

Operating model

The same cycle at every organisation I've led. The step most teams skip is the last arrow — the one that carries incident learnings and cost signals back into the next planning round instead of into a document nobody reopens.

  1. 01

    Direction

    • Business and compliance goals
    • OKRs and quarterly roadmap

  2. 02

    Design

    • Architecture guardrails
    • Team design and staffing

  3. 03

    Delivery

    • Pods shipping independently
    • FinOps: cost per transaction

  4. 04

    Feedback

    • SLIs, SLOs and error budgets
    • Incident review and RCA
    • Tech-debt backlog

Feedback returns to direction — the cycle repeats

Experience

Sixteen years, seven companies

Telecom to marketplaces to payments to enterprise SaaS. The scope changed; the habit of owning the whole system rather than a slice of it didn't.

  1. Jul 2026 — Present

    Hyderabad, India — Remote

    Current

    Director, Platform Engineering

    AssetMark

    Leading platform engineering for a US wealth-management technology firm, with ownership of shared services, developer experience and the engineering standards the product teams build on.

  2. Mar 2025 — Apr 2026

    Bengaluru, India

    Senior Engineering Manager, Card Payments

    Razorpay

    Owned two payment platforms — card recurring and e-mandate, and partner pay programmes — running 5–15K TPS and $60–90M in daily volume.

    • Led 21 engineers across three pods and influenced roughly 33 more across routing, terminal and cross-border payment domains.
    • Drove the re-architecture of a legacy PHP API monolith into Go services for card and recurring payments, against the internal security, compliance, reliability, efficiency and delivery bar.
    • Closed a visibility gap where merchant-impacting failures surfaced 20 minutes late: a Kafka, Spark Streaming and Apache Pinot pipeline brought detection under one second for 40+ high-value merchants.
    • Shipped Cred Pay and Apple Pay as the year's primary OKRs, and cut logging cost per transaction from ₹0.09 to ₹0.03.

  3. Jun 2023 — Feb 2025

    Pune, India

    Director, Software Engineering

    Guidepoint

    Ran the India engineering organisation for CRM modernisation, advisor discovery, client portals and AI-assisted products, partnering directly with product leadership in New York.

    • Delivered a full PHP 8 migration of the ecosystem in two months, unblocking SOC 2 compliance and continuity for enterprise clients.
    • Decomposed monolithic systems into services and wired Databricks ETL into the advisor discovery platform.
    • Designed a federated micro-frontend architecture unifying Advisor Search, Insights 2.0 and the AI co-pilot, with LLM-backed advisor matching.
    • Owned the org design and stakeholder plan with the CTO that turned the Pune office into a Global Capability Center.

  4. Feb 2021 — Jun 2023

    Pune, India — Remote

    Manager, Engineering

    Acquia

    Built and led the Cloud Next upgrades and data platform teams across IST and Boston hours, spanning junior through principal engineers.

    • Delivered $660K in annual COGS savings by migrating customers from Cloud Classic to Cloud Next on a multi-tenant Aurora MySQL model.
    • Owned the Acquia Cloud Next roadmap features that shortened enterprise onboarding.
    • Pushed Kubernetes, Terraform, CI/CD automation and SRE practice across the cloud and digital asset management platforms.

  5. Nov 2016 — Feb 2021

    Gurugram, India

    Technical Lead to Engineering Manager

    MobiKwik

    Took insurance, lending and travel lines from nothing to production — MVP scope, partner integrations, architecture and delivery.

    • Built the insurance cross-sell framework (Stargate) that allowed multiple service bundles in a single transaction.
    • Led the shift from wallet-only to wallet-as-payment-gateway in line with RBI PPI regulation.
    • Held 99.90% availability on recharge and bill payments with circuit-breaker isolation, and sub-200ms API latency across Euronet, BillDesk and BillAvenue.
    • Grew teams from zero to twelve engineers, and lifted recharge GMV from ₹181 Cr to ₹210 Cr a month through a targeted tech-debt programme.

  6. Sep 2015 — Nov 2016

    Gurugram, India

    Staff Software Engineer

    ShopClues

    Led six engineers across catalogue ingestion, moderation and inventory.

    • Raised bulk ingestion from 30K to 100K products a day and cut banned-product enforcement from five days to under 24 hours.

  7. Nov 2011 — May 2015

    Noida, India

    Senior Software Engineer to Lead Engineer

    InfoEdge — Naukri.com

    Built the Background Check and Priority Applicant platforms, and migrated legacy systems from Symfony 1.x to 2.x.

    • Moved the CRM email and SMS pipeline SLA from 3 PM to 11 AM for over a million active users, and introduced Git, Docker and agile practice across the teams.

  8. Dec 2009 — Mar 2011

    Gurugram, India

    Software Engineer

    Comviva Technologies

    Module developer on PreTUPS, a prepaid top-up platform deployed across 50+ telecom operators, and the on-site point of contact for Vodafone East, West and North migrations.

Education

B.Tech, Computer Engineering — Jamia Millia Islamia, New Delhi

2005 — 2009 · First Division

Outcomes

Numbers I'm accountable for

Every line below had a cost centre, an owner and a date. They are the results I'd bring up in a reference call.

  • < 1s

    Failure detection on payment flows

    Merchant-impacting issues used to surface 20 minutes late. A Kafka → Spark Streaming → Apache Pinot pipeline brought it under a second for 40+ high-value merchants.

    Razorpay

  • $660K

    Annual COGS removed

    Migrating customers from Cloud Classic to Cloud Next on a multi-tenant Aurora MySQL model, without a customer-visible cutover event.

    Acquia

  • ₹0.09 → ₹0.03

    Logging cost per transaction

    An org-wide tech-debt and FinOps programme that made observability spend proportional to the value of what was being observed.

    Razorpay

  • 2 months

    Full PHP 8 migration

    The entire product ecosystem moved in a single quarter, unblocking SOC 2 compliance and continuity for enterprise clients.

    Guidepoint

  • 99.90%

    Availability on recharge and bill payments

    Circuit-breaker isolation across Euronet, BillDesk and BillAvenue as official BBPOUs, holding sub-200ms API latency.

    MobiKwik

  • ₹181 → ₹210 Cr

    Monthly recharge GMV

    Earned by fixing the top operational pain points rather than adding features — a tech-debt initiative with a revenue line attached.

    MobiKwik

  • 30K → 100K

    Products ingested per day

    Catalogue ingestion rebuilt on REST services, with banned-product enforcement cut from five days to under 24 hours.

    ShopClues

  • 40+

    Engineers hired since 2016

    Across juniors to principals, in-house and contract, in India and the US — including two teams taken from zero to twelve.

    Across roles

Also shipped

  • Cred Pay
  • Apple Pay
  • Google Pay
  • RBI PPI compliance
  • SOC 2 readiness
  • PCI-DSS environments

Toolkit

What I've worked with

Listed without proficiency percentages, because nobody can honestly assign one. These are the tools I've made production decisions with, and can still be useful in a design review about.

Languages
JavaGoPHPPythonShell
Frameworks
Spring BootEchogRPCSymfony / API PlatformHibernate
Architecture
MicroservicesEvent-driven systemsCQRSTransactional outboxCircuit breakerStrangler fig migrations
Cloud & delivery
AWSAzure AKSKubernetesDockerTerraformAnsibleArgo CDGitHub ActionsJenkins
Data & messaging
MySQLPostgreSQLTiDBMongoDBCassandraRedisKafkaRabbitMQAmazon SQSApache PinotElasticsearchDatabricks
Observability & FinOps
OpenTelemetryDatadogGrafanaPrometheusELKMicrometerZipkinPagerDutyCloudZero
AI-assisted engineering
Claude CodeCursorChatGPTNotebookLMn8n workflows
Leadership
Org designHiring & coachingOKR planningStakeholder managementIncident managementCost optimisation

Currently exploring

RAG pipelines, LLM agents, the Model Context Protocol, and n8n for workflow automation.

Contact

Get in touch

I answer email properly — a short note about the problem you're working on gets a better reply than a connection request. Currently at AssetMark, and always happy to talk platform or payments work, including opportunities across India, Singapore and the UAE.

bharatv@outlook.in
Based in
Pune, Maharashtra, India
Languages
English, Hindi
Elsewhere
LinkedInTwitterGitHubMediumDev.toHashnodeStack Overflow

Worth a conversation about

Platform strategy
What a platform team should own, expose and refuse to own.
Payment systems
Card, recurring and e-mandate flows, partner failover, reconciliation, PCI-DSS.
Monolith decomposition
Strangler-fig migrations that finish, rather than leaving two systems to maintain.
Org design and GCC build-out
Standing up an India centre with real ownership, and running IST–EST delivery.
Reliability and FinOps
On-call that people can live with, and observability spend that stays proportionate.
Moving into leadership
Coaching, interview preparation and resume reviews for engineers making the jump.