Skip to content

Work I've delivered, and what I can build for you

I partner with business owners and engineering teams to deliver production software end to end, either building it myself or consulting alongside your engineers. You work directly with the engineer building your system, get clear progress updates at every stage, and receive well-documented, maintainable software your team can run and extend.

Complex problems I'm hired to solve

The work where experience matters most. Each one is backed by a project I've delivered; open any of them to see how I approach it.

And much more, shaped around your requirements: web applications, APIs, internal tools, business websites and anything in between. Describe your project and I'll recommend the right approach.

AWS Governance and Security

As AWS accounts multiply, security drifts. Policies exist on paper, monitoring covers only part of the estate, and long-lived keys pile up. Fixing it on a live estate means changing the rules under engineers who deploy every day.

What it covers

  • Landing zone setup with AWS Control Tower over existing accounts
  • Preventive guardrails: service control, resource control and declarative policies
  • Security monitoring with Security Hub and the CIS benchmark across every Region in use
  • Identity clean-up: IAM Identity Center, stale access keys and risky permission sets
  • CI/CD access through GitHub OIDC instead of stored keys
  • Moving accounts safely between AWS organizations

What usually goes wrong

  • Guardrails that look enabled but evaluate nothing, or cover one service and miss another
  • Organization-wide policies applied in one step, breaking production pipelines
  • Security findings trusted as complete when monitoring covers only part of the estate

How I approach it

  1. 1Map every account, identity and existing policy before changing anything
  2. 2Roll guardrails out from a test unit outward, verifying each stage
  3. 3Turn on estate-wide security monitoring and confirm it is actually evaluating
  4. 4Hand over runbooks and catalogues your team can run without me

Where I've done it

Usually 2 to 8 weeks, as a fixed-scope project or a consulting engagement. Every engagement starts with a free intro call.

Cloud Migrations

A migration is only as safe as its riskiest dependency: the databases, the encryption keys tied to the old cloud, and the addresses customers already use. Moving the applications is the easy part.

What it covers

  • Kubernetes workloads moved onto Amazon EKS
  • Databases and object storage moved onto Amazon RDS and S3
  • Encrypted data re-encrypted for the new platform and verified
  • Secrets, service mesh and GitOps tooling: Vault, Istio and Argo CD
  • Infrastructure as code with Terraform
  • Redirects so existing links keep working after cutover

What usually goes wrong

  • Encrypted data moved without the keys that can read it
  • Cutovers planned around applications rather than the data and dependencies beneath them
  • Old addresses switched off, breaking links customers rely on

How I approach it

  1. 1Inventory every component and decide what moves, what is rebuilt and what retires
  2. 2Rehearse the data move, including re-encryption, and verify it record by record
  3. 3Cut over once, on a planned date, with redirects and a way back
  4. 4Leave an operations guide for the team that runs the new platform

Where I've done it

Usually 4 to 12 weeks, scoped after a short assessment. Every engagement starts with a free intro call.

AI Agents and Assistants

CCDV-F Certified

Getting a model to answer is easy. Making it act safely on real systems, respond quickly and behave predictably is where most AI features stall.

What it covers

  • Tool-using agents that act on real accounts through OAuth
  • Real-time voice and streaming responses
  • Latency measured and reduced where users feel it
  • Automated tests that cover agent behaviour
  • Backend services and native mobile apps around the AI

What usually goes wrong

  • Demos that break once the AI touches real accounts and real data
  • Slow responses that make an assistant feel unusable
  • No automated tests, so every prompt change is a gamble

How I approach it

  1. 1Define the jobs the AI must do and the tools it may use, with clear permissions
  2. 2Build tool integrations with authentication and auditing in mind
  3. 3Measure latency and cut it where users feel it
  4. 4Cover behaviour with automated tests before it ships

Where I've done it

Usually 2 to 6 weeks for a first production feature, then ongoing iterations. Every engagement starts with a free intro call.

ETL Pipelines and Data Synchronization

When data lives in several systems, it has to be extracted, transformed and kept in step while every side keeps changing. Pipelines that work on day one often fail quietly by day ninety.

What it covers

  • Extract, transform and load pipelines between business systems
  • One-way and two-way synchronization with a clear system of record
  • Queue-based workers with retries and a ledger of every change
  • Reconciliation that finds and repairs drift
  • Phased historical migrations with dry runs
  • Integrations with SaaS APIs such as Jira Cloud

What usually goes wrong

  • Duplicates and update loops when systems write to each other
  • Silent drift that nobody notices until a customer does
  • Historical backfills run straight into production without a rehearsal

How I approach it

  1. 1Agree which system owns each piece of data
  2. 2Build idempotent pipelines with queues, retries and a ledger of what moved
  3. 3Add reconciliation that detects and repairs drift automatically
  4. 4Load historical data in phases, with dry runs first

Where I've done it

Usually 3 to 8 weeks to build, with an optional retainer to operate it. Every engagement starts with a free intro call.

Microservices, Distributed Systems and APIs

When a platform is split into many services, the hard problems move into the space between them: events that arrive twice or out of order, balances that must reconcile to the cent, and partner integrations that fail on their own schedule.

What it covers

  • Betting engines with fund reservations, settlement and accounting
  • Commission and accounting logic that reconciles across services
  • Event-driven messaging with Kafka, Amazon SQS, SNS and EventBridge
  • Third-party casino and odds-feed integrations
  • Clinical-trial recruitment engines matching patients with investigators, with Solr search
  • Contract and e-signature workflow management running as microservices

What usually goes wrong

  • Events processed twice or out of order, corrupting balances and state
  • Services that disagree about shared data, so users see what should not exist
  • Partner integrations without retries, timeouts or idempotency

How I approach it

  1. 1Map the money and data flows end to end, across every service
  2. 2Design APIs and events to be idempotent and safe to retry
  3. 3Build consistency checks and reconciliation into the system itself
  4. 4Fix production issues at the root and verify against real data

Where I've done it

Hourly, as a retainer, or as a fixed-scope build after a short design phase. Every engagement starts with a free intro call.

End-to-End Product Development

Shipping version one is half the job. A product also needs payments, search visibility, performance and someone who keeps it running once real users arrive.

What it covers

  • Product shaping: users, core flow and launch scope
  • Full-stack web applications with online payments
  • Performance and Core Web Vitals from day one
  • Search visibility and structured data
  • Running the product after launch

What usually goes wrong

  • A launch with no plan for monitoring, updates or support
  • Performance and search visibility left until after launch
  • Several vendors, and nobody owning the whole

How I approach it

  1. 1Shape the product with you: the users, the core flow and what launches first
  2. 2Build the full stack: design, backend, payments and infrastructure
  3. 3Launch with performance and search visibility already in place
  4. 4Keep running and improving it after launch

Where I've done it

A fixed-scope version one, then a retainer. Every engagement starts with a free intro call.

How a project runs

  1. Step 1

    Project Induction

    We discuss your goals, users, constraints and timeline, and agree on what success looks like.

  2. Step 2

    Requirements and Planning

    Product requirements, technical approach, milestones and estimates, captured in a clear written plan.

  3. Step 3

    Architecture and Design

    System design, data models, security and infrastructure decisions, reviewed with you before development starts.

  4. Step 4

    Development and Delivery

    Iterative development with CI/CD, code reviews and working previews at every milestone.

  5. Step 5

    Launch and Handover

    Secure production deployment, observability, documentation and a walkthrough for your team.

  6. Step 6

    Maintenance and Continuous Support

    Monitoring, updates, security patches and ongoing improvements as your product grows.

Ways to work together

Choose the option that fits, or describe the work and I'll recommend the right model. Every engagement starts with a free intro call.

Build work

From $50 per hour

Web applications, APIs, integrations, AI features and cloud infrastructure, billed by the hour.

  • Weekly progress updates
  • You own the code, documentation and access
  • Hours that scale with your needs
Get a quote

Consulting and reviews

From $75 per hour

Architecture and code reviews, AWS governance and security, and migration planning.

  • Written findings with prioritised recommendations
  • Working sessions with your team
  • Suited to short, focused engagements
Get a quote

Monthly retainer

$45 per hour, minimum 20 hours a month

Ongoing development, maintenance and support at a reduced rate, with capacity reserved for you each month.

  • Priority response
  • Monitoring, updates and security patches
  • A monthly summary of work and hours
Get a quote

Fixed-scope project

Quoted after a free intro call

A defined scope, clear milestones and a single agreed price, for work that can be specified up front.

  • Written scope and milestones
  • Milestone-based payments
  • Scope changes agreed in writing before work begins
Get a quote
Working hours
Remote from India (IST, UTC+5:30). Live calls are scheduled by arrangement, with overlap for the UAE, Europe and the US East Coast, and written progress updates in between.
Pricing and invoices
All rates are in US dollars, and invoices can be issued in USD, EUR, AED or INR. Clients based in India can request pricing in rupees.

Case studies

My own product, 2026

RajyaPrep, a daily practice platform for competitive exams

A paid daily practice platform for UPSC, State PSC, SSC, banking, railways and defence aspirants. I designed, built, launched and operate the whole product myself.

  • A 30-question daily paper, with every option explained and the source of every question named
  • Subject tests across eleven subjects, mapped to more than thirty named exams
  • One-time plans from a single paper to a full year, with no auto-renewal

Product DesignFull-Stack Web ApplicationAI-Assisted ContentOnline PaymentsSEOCore Web VitalsMobile-First Design

My own product, 2026

Zuni, a real-time voice AI assistant

A personal voice AI assistant I built end to end: a native iPhone app with natural, interruptible conversation and a backend that lets the AI act on real accounts.

  • Version 1.0 shipped on both the iPhone app and the backend in about two weeks of build time
  • Full-duplex conversation: it listens while it speaks and can be interrupted mid-reply
  • Model time to first token cut from about 3.4 seconds to under 1

Native iOSReal-Time VoiceStreaming SpeechLLM Tool UseOAuthPythonAutomated Testing

Cybersecurity managed services provider, 2026

AWS governance and security for a cybersecurity provider

For a cybersecurity managed services provider, I set up AWS Control Tower over a live multi-account estate, rolled out preventive guardrails and CIS security monitoring, cleaned up identity and access, and built a safe method for moving accounts between AWS organizations.

  • AWS Control Tower set up over a live estate, with every account the client controls in its main organization enrolled
  • Baseline guardrail policies on nearly every enrolled account, rolled out from a test unit outward with no production incidents
  • CIS security monitoring across the centrally managed workload accounts in every Region in use

AWS Control TowerAWS OrganizationsService Control PoliciesResource Control PoliciesDeclarative PoliciesIAM Identity CenterAWS Security HubCIS BenchmarkAWS ConfigIAM Access AnalyzerCloudTrailCloudFormation StackSetsTerraformGitHub OIDCPython

Cybersecurity company, 2026

Moving a production platform from GCP to AWS

For a cybersecurity company, I led the move of several production applications from Google Kubernetes Engine to Amazon EKS, including databases, object storage and encrypted customer data.

  • Five applications live on AWS after a single planned cutover, and a sixth retired instead of moved
  • Encrypted customer data re-encrypted for the new platform and verified field by field
  • Old addresses redirected so existing links kept working

KubernetesAmazon EKSTerraformAmazon RDS for PostgreSQLAmazon S3HashiCorp VaultArgo WorkflowsGitHub ActionsDjangoPython

Cybersecurity managed services provider, 2026

Two-way sync between a client portal and Jira

For a cybersecurity managed services provider, I built the service that keeps their customer portal and Jira Cloud in step, in both directions.

  • Customer, work-item, comment and assignment records kept in sync automatically
  • Historical records migrated in phases, with dry runs first
  • A reconciliation engine that finds and repairs drift between the two systems

PythonFastAPIPostgreSQLAmazon SQSAPSchedulerJira Cloud APIKubernetesAmazon EKSGitHub Actions

Multi-brand digital awards platform, 2026

Maintaining a multi-brand awards platform

For a company running several digital awards programs on one shared platform, I support and fix a mature .NET application that handles entries, judging, voting and winners.

  • A cloud development environment that runs the full platform safely against staging
  • A verified guide to the data model and the entry-to-winner workflow
  • Fixes checked before and after against database ground truth

ASP.NET MVC.NET FrameworkEntity FrameworkSQL ServerAngularJSAWS EC2Amazon RDS

Online gaming and sports platform, 2025

Online Gaming and Sports Platform

As one of the core backend engineers on an online gaming and sports platform of about ten microservices, I designed and owned the player-facing market and bet-placement service in Go, and contributed across the rest of the platform, from the betting engine to the pipeline that turns every bet into live profit and loss for partners.

  • Designed and owned the player-facing market and bet-placement service, written in Go
  • Fund reservations serialized per player and market with a distributed lock, so concurrent bets do not race for the same balance
  • Every accepted bet turned into live exposure and profit and loss up the partner hierarchy

GoFiberPythonMicroservicesApache KafkaRedisPostgreSQLAmazon SQSAmazon SNSAWS EventBridgeOpenTelemetryDockerAmazon ECR

Clinical trial recruitment software company, 2022

Multi-tenant clinical trial recruitment platform

At a clinical trial recruitment software company, I worked across the full stack of one Django platform serving several leading US clinical research institutions, each with its own database, features and branding.

  • One codebase served several leading US clinical research institutions, each with its own database, features and branding
  • Trial listings imported and refreshed from ClinicalTrials.gov by NCT ID
  • Matching that weighed every facet of a trial, including condition, age, gender and location

PythonDjangoDjango templatesJavaScriptjQueryReactPostgreSQLApache SolrApache HTTP ServerClinicalTrials.govGmail

Have a project in mind?

Share your goals, timeline and any constraints, whether you need it built or want expert advice. I read every message myself, as the engineer who would build it, and reply with a clear view on approach, scope and next steps.

Hire me