Logesh Thulasiraman | Senior Full-Stack Engineer

Nº 01 — About

Fig. 1 The author.

Hi, I'm Logesh — I build things that live on the internet.

I like owning them from the first sketch to the 2am production fix. Eleven-plus years, six industries: hospital EHRs, real-time chat, real estate apps, retail analytics, insurance APIs and enterprise SaaS. Different domains, same itch — take a messy problem and ship something people actually use.

Right now I'm hooked on AI: what happens when LLMs stop being demos and start working inside real products, on real data, under real load.

I believe learning grows by sharing, so I write about what I build in the journal. Call me an explorer in the tech world, always one new tool away from a side project.

Away from the keyboard: running, chasing new places and flying a drone over them. The proof is in chapter six.

“Pixel to API to AI.”
— the job, in five words

i.
Users first, stack second.
Start from the experience, then pick the simplest tech that delivers it.
ii.
AI is a feature, not magic.
It gets tests, fallbacks and monitoring like any other code path.
iii.
Own it in prod.
If I build it, I can debug it at 2am from logs and traces alone.

Nº 02 — Experience

Where I've shipped.

CoverForce Full-Stack Engineer InsurTech

Unified API to quote & bind commercial insurance

Engineered parts of a unified API that lets brokers and platforms generate, compare and bind commercial insurance quotes across 20+ carriers in one call.
Built carrier integrations: adapters that translate very different carrier REST, SOAP and XML APIs into one normalized quote schema, with timeouts, retries and idempotent submissions.
Applied LLMs to insurance submissions, extracting structured data from applications and documents, with confidence thresholds and a human-review fallback.
Contributed to the white-label agent portal on top of the API, from React screens down to the services behind them.

PythonNode.jsReactREST & SOAPLLMsAWS

NIQ Principal Engineer Consumer intelligence

The Full View of what the world buys

Helped power NIQ's retail measurement, the data that tells the world's biggest brands and retailers what consumers buy, where and why, across 90+ countries and over $7 trillion in consumer spend.
Built large-scale pipelines on Azure (Data Lake, U-SQL) that turn raw point-of-sale data from thousands of stores into market-share and sales-performance insights.
Designed a custom Airflow platform on Kubernetes whose workers scale up and down with demand, absorbing heavy reporting cycles without paying for idle compute.
Built Python APIs and React dashboards that put FMCG sales trends in front of client teams, and grew from Software Engineer to Principal Engineer, setting technical direction for the team.

PythonAirflowKubernetesAzure Data LakeU-SQLReact

Xome Software Engineer PropTech

Online real estate marketplace & auctions

Architected and shipped Xome Pro, a CRM app for real estate agents, to the App Store and Play Store from a single React Native codebase.
Built Node.js microservices for lead routing, agent assignment and automated follow-ups, with push notifications to field agents.
Integrated the app with Xome's listing and relationship-management systems so agents, buyers, sellers and RM teams worked from the same data.

React NativeNode.jsMicroservicesPythonMongoDB

BroadSoft (Cisco) Software Developer Unified comms

Team-One collaboration app

Built real-time messaging for Team-One, a Slack-style collaboration app: WebSocket channels, presence and message sync across devices.
Worked on in-app video streaming and calling, alongside BroadSoft's cloud communications platform.
Shipped web, mobile and Electron desktop clients for Windows and macOS from a shared JavaScript codebase.
Proposed and built an extension marketplace with public APIs so customers and partners could build their own integrations.

Node.jsJavaWebSocketsReactElectron

athenahealth Software Developer HealthTech

Cloud EHR for hospitals

Built features for athenaClinicals for Hospitals, an inpatient EHR covering patient charts and medication administration during a hospital stay.
Built IoT integrations that capture patient vitals from bedside devices straight into the chart, removing manual entry.
Created an automated end-to-end testing framework for EHR products that became the safety net for every release.
Set up CI/CD pipelines that ran the test suites on every change.

PerlRubyNode.jsIoTCI/CD

Education — B.E. Electronics & Communication Engineering, Anna University, Chennai.

Nº 03 — Approach

How I build products, in six movements.

I
Design
Start from the person using it. A rough flow and a clear contract between screens and services saves weeks later.
II
Interface
Fast, accessible screens on web, mobile and desktop. Loading, empty and error states get designed, not discovered.
III
API
APIs are products too. Clear contracts, predictable errors and services that stay up when a dependency doesn't.
IV
Data
Pick the store for the access pattern. Model it so the common queries are cheap and migrations are boring.
V
AI
LLMs become useful when they are wired into real data and real actions, and treated like any other dependency that can fail.
VI
Ship
Ship small and often, and know within minutes whether it worked. If I built it, I can debug it in production.

Nº 04 — Index of tools

Frontend & mobile

React
React Native
TypeScript
Redux
Electron

Backend

Java
Spring Boot
Node.js
Python
PostgreSQL
MongoDB

AI

OpenAI
Claude
Cursor
Copilot
RRAG
{}Agents

Cloud & delivery

AWS
Azure
Docker
Kubernetes
CI/CD

Nº 05 — Selected work

Features & case notes.

api.coverforce.com

              POST /v1/quotes
              { "business": "Joe's Bakery", "lines": ["WC", "BOP"], "carriers": "all" }
              200 OK · 20 carriers queried · 1.8s
              {
                "quote_id": "q_7f3c9a",
                "quotes": [
                  { "carrier": "Carrier A", "premium": 4120, "bindable": true },
                  { "carrier": "Carrier B", "premium": 4385 },
                  { "carrier": "Carrier C", "premium": 4610 },
                  { "carrier": "Carrier D", "premium": 5040 }
                ]
              }

CoverForce · InsurTech

One API, every carrier

A unified API that turns one request into real-time commercial insurance quotes from 20+ carriers, ready to compare and bind. I built carrier adapters that tame wildly different REST, SOAP and XML APIs, plus LLM-powered extraction that reads submission documents so brokers don't have to retype them.

PythonNode.jsReactLLMsAWS

Retail measurement · Salty snacks

Market share Last 52 weeks All channels

Value sales$1.28B▲ 4.1%

Brand A share31.0%▲ 0.6 pts

Stores tracked14.2Kweekly

Brand ACategory avg.

Brand A31%

Brand B24%

Brand C18%

Brand D12%

Private label9%

NIQ · Consumer intelligence

Retail measurement at scale

The pipelines behind NIQ's retail measurement: point-of-sale data from thousands of stores, cleaned and modelled on Azure into the market-share numbers FMCG brands plan around. Orchestrated by a custom Airflow on Kubernetes that scales with the reporting cycle, and surfaced through React dashboards.

PythonAirflowKubernetesAzure Data LakeReact