About Suraj Mittal

Suraj Mittal is an Engineering Manager based in Bengaluru, India, with over ten years of experience spanning platform engineering, data systems, and fintech infrastructure. He currently leads a five-person platform backend team at Setu, where his work covers reliability engineering, compliance, developer experience, AI tooling, and cloud cost control. His career arc moves from shipped code to shipped teams, with stops at Kirusa, Freecharge, Hashworks, EdGE Networks, Rivi, Artefact, and Setu. He has built Kubernetes platforms, billing systems, ClickHouse pipelines, hiring loops, on-call tooling, and team rituals that make engineers better at their craft. His technical range includes Go, Python, Ruby, Kubernetes, AWS, Terraform, PostgreSQL, ClickHouse, Kafka, and distributed systems. He believes that strong teams make better systems than heroic individuals, and that the useful question is rarely what can we build but what should become easier for the people depending on us. He is interested in engineering leadership roles where platform thinking, people development, reliability, and product judgment all matter. You can reach him via email at me@suraj-mittal.dev or connect on LinkedIn and GitHub.

Notable achievements include reducing billing close time from fifteen days to under one day by building a unified billing platform at Setu, cutting centralized logging costs from thirty-five thousand dollars to five thousand dollars monthly through a Vector, ClickHouse, and Warpstream redesign, and improving ClickHouse analytics performance by ten times with global caching. He shipped an AI code review bot that achieved over ninety percent developer satisfaction, rebuilt hiring processes with bug bash and live coding rounds later adopted across the organization, and coached two engineers to exceptional performance ratings. Earlier work includes building hotel recommendation systems and route optimization at Rivi, improving skills coverage to ninety percent through knowledge graphs at EdGE Networks, and creating ranking systems that matched candidates to opportunities at Hashworks. His open source work includes the OpenCode IPython Plugin and Magic Context, tools for AI coding agents with persistent kernels and cross-session memory management.

Issue Jun 2026Bengaluru · Engineering Manager

Suraj Mittal

Start a conversation

I build teams that enjoy hard problems, and systems that survive real traffic.

Engineering manager with 10+ years across platform engineering, data systems, and fintech infrastructure. I lead a five-person platform backend team at Setu, where the work spans reliability, compliance, developer experience, AI tooling, and cost control. My range is unusual by design: I have shipped Kubernetes platforms, billing systems, ClickHouse pipelines, hiring loops, on-call tooling, and team rituals that make engineers better at their craft.

Operating question“The useful question is rarely 'what can we build?' It is 'what should become easier for the people depending on us?' That question has shaped how I lead teams, choose systems, and decide what deserves attention.”

Operating style

The work is people, systems, and the space between them.

I like leadership that is specific enough to be useful. These are the habits I bring into platform teams, incident reviews, hiring loops, architecture conversations, and the messy middle where engineering choices become business outcomes.

No. 01

Curiosity beats certainty

I ask before I prescribe. The best technical decisions usually appear after enough people explain what hurts, what is expensive, and what they are quietly working around.

No. 02

Systems need editors

Good engineering leadership is often subtraction: fewer unclear owners, fewer fragile paths, fewer meetings that only exist because the tooling is bad.

No. 03

Impact should be visible

The work has to change something measurable: a billing close, a cloud bill, an on-call handoff, a hiring signal, a developer's day.

No. 04

People scale the architecture

Strong teams make better systems than heroic individuals. My job is to grow judgment, ownership, and enough psychological safety for useful disagreement.

If that sounds like the kind of engineering culture you want to build, we will probably have a useful conversation.

Career arc

From shipped code to shipped teams.

The throughline is not a single stack. It is a habit of moving closer to the hard operating problems: reliability, scale, cost, hiring, ownership, and decision quality.

201601

First shipped software

Started at Kirusa building Instavoice for Windows 10 and an Alexa skill, learning how real users find the edges of every product.

201702

Fintech at product speed

Joined Freecharge and got a first look at financial products where reliability, polish, and iteration speed all matter at once.

2017-1803

Machine learning in the workflow

Built ranking and recommendation systems at Hashworks, matching candidates to opportunities with practical NLP and deep learning.

2018-1904

Data science with structure

At EdGE Networks, worked on knowledge graphs, entity extraction, and models that improved skills and roles coverage to 90%.

2019-2105

Architecture met leadership

Owned backend architecture at Rivi, built recommendation systems, led hiring across teams, and worked directly with enterprise clients.

2021-2206

Solutions and advisory

Led analytics delivery for Nissan through Artefact while advising Rivi on technical direction, hiring, and platform choices.

2022-2407

Platform foundations

Built Kubernetes, GitOps, auth, logging, and real-time data infrastructure at Setu for teams operating in regulated fintech.

2024-2508

Staff-level systems

Shipped billing and data platforms, cut billing close from 15 days to under 1, and improved analytics performance by 10x.

2025+09

Engineering management

Now leading platform engineers, rebuilding hiring, rolling out AI tooling, reducing cloud spend, and raising team ownership.

Proof points

A few numbers that point to the shape of the work: operational leverage, better developer experience, and teams that move with more confidence.

10+
years across systems, data, and leadership
5
platform engineers and interns managed directly
$30k+
monthly infrastructure savings from logging redesign
90%+
developer satisfaction on internal AI review tooling

Selected builds

Experiments with teeth.

Earlier research and engineering work across optimization, representation learning, and routing protocols. Different domains, same appetite for making abstract systems concrete.

Particle Swarm Optimization for LSTMs

Explored whether particle swarm optimization could tune constrained-boundary LSTMs faster than the default training path, with measurable gains in model performance and training time.

PSOLSTMDeep Learning
Sparse Document Composition Vectors

Built an unsupervised document representation method using soft clustering, word2vec, and thresholded composition vectors for downstream NLP tasks.

NLPUnsupervised Learningword2vec
Threshold-Sensitive Multihop Routing

Designed a self-aware routing protocol for battery-limited sensor networks and evaluated it against common approaches under constrained conditions.

Wireless Sensor NetworksRoutingProtocols
OpenCode IPython Plugin

Python tool for OpenCode with persistent IPython kernels, rich output support, and 50+ preloaded helper functions. Manages isolated Python environments with LRU session pooling and WebSocket-based Jupyter kernel communication.

TypeScriptPythonJupyterWebSocket
Magic Context

Fork of cortexkit/magic-context adapted for oh-my-pi. Cache-aware infinite context plugin for AI coding agents with background historian, cross-session memory, and overnight dreamer for consolidating project knowledge.

TypeScriptSQLiteEmbeddingsLLMoh-my-pi

Leadership record

Where the operating system changed.

A decade across fintech, data science, and platform engineering. Open each entry for the systems shipped, teams shaped, and leverage created.

Engineering Manager

April 2025 - Present

Lead a four-engineer platform backend team plus one intern across reliability, compliance, AI tooling, and developer productivity. Rebuilt hiring with bug bash, live coding, and technical curiosity rounds that were later adopted org-wide. Coached two engineers to exceptional performance ratings. Shipped an AI code review bot with 90%+ developer satisfaction, cut recurring support tasks from five minutes to under ten seconds, migrated Kafka to Warpstream to halve spend, and delivered regulated fintech work including CERSAI CKYC, UPI AML services, platform audits, Karta AI onboarding, n8n workflow enablement, and Slack-to-Jira automation.

ManagementInfrastructurePythonCompliance

Staff Software Engineer

April 2024 - March 2025

Built the unified billing platform for prepaid, postpaid, percentage, and volumetric models, reducing billing close from 15 days to under one. Reworked centralized logging with Vector, ClickHouse, Warpstream, and S3 to reduce monthly spend from $35k to $5k. Shipped real-time S3-to-ClickHouse ingestion, improved ClickHouse analytics performance by 10x with global caching, introduced Karpenter autoscaling, and solved Docker Hub pull limits with a Go pull-through cache backed by ECR.

RubyPythonGoClickHouseSupersetPostgresDistributed Systems

Senior Software Engineer

March 2022 - March 2024

Built Kubernetes platform foundations with GitOps, federated deployments, multi-cluster management, standardized Helm charts, ALB provisioning, external secrets, and Cognito auth. Designed central auth with Keycloak, maintained Temporal for durable workflows across teams, and built CDC pipelines from Debezium through Kafka and Warpstream into ClickHouse.

GoAWSPostgresKubernetesHelmGitOps

Technical range

Capability clusters, not keyword soup.

The connective tissue is platform judgment: what to automate, what to simplify, and what to make easier for the next engineer.

Leadership

Set 01
People ManagementTeam BuildingHiring & Performance1:1sCareer DevelopmentStakeholder ManagementRoadmap Planning

Platform & Ops

Set 02
Platform EngineeringOn-call & Incident ResponseCompliance & AuditsData Platforms

Technical

Set 03
PythonGoRubyKubernetesAWSTerraformCI/CDGitOpsObservability

Data

Set 04
PostgreSQLClickHouseKafkaDistributed Systems

Credentials

Formal proof, practical bias.

The degree and certifications matter. The bigger education came from production systems, sharp teammates, and incidents that made the architecture more honest.

Education

Bachelor's Degree, Computer Science

Visvesvaraya Technological University

2012 - 2016

Recognition

Contact

Let’s compare notes on teams, systems, and leverage.

I am interested in engineering leadership roles where platform thinking, people development, reliability, and product judgment all matter. If you are building that kind of team, reach out.

ColophonReact, TanStack Router, Framer Motion, OKLCH, and an editorial bias
© 2026 Suraj Mittal