KodeLogiK designs and builds AI, data and enterprise software for companies across the United States, from our base in Pune, India. Senior engineers, production-grade work, and a founder who stays in the room from the first call to the last release.
AI, machine learning, data, application, SaaS and e-commerce development: six practices, one team. Most clients start in one and grow into the others as the system matures.
Agentic AI, assistants and automation that run in production, not in a demo.
Models trained on your own order history to forecast demand by day, season, region and event.
One trusted source for every number in the business, and the reports that run on it.
Web, mobile and cloud applications built on .NET and deployed on AWS or Azure.
From idea to a subscription product that is ready to launch, scale and bill.
Online stores and back-office systems that keep billing, inventory and operations in step.
Software we have designed, built and run for clients in the United States. Client names are withheld under agreement and shared on request; figures are the client’s own.
Cabinet doors are a considered purchase and shoppers could not picture a door style in their own kitchen before ordering. The retailer’s long-running nopCommerce platform also needed steady engineering alongside any new feature.
A generative visualiser on OpenAI GPT-4o that renders the customer’s kitchen with the chosen door style, built into the existing storefront rather than a separate tool, plus ongoing platform work: shipping API migrations, email marketing integrations and database optimisation.
Shoppers now see their own kitchen in the door style they are considering, and more of them get in touch and order. The same multi-year engagement keeps the platform fast through peak sale campaigns.
Years of order history sat in the store’s database, but planning still worked from last year’s totals. Demand moves differently by region and by season, and events like Black Friday distort the averages for everyone.
A forecasting model in ML.NET over the store’s order history that predicts demand day by day and month by month, with separate components for festival and sale peaks, for each region, and for climate seasons by region.
Forecasts the retailer can plan stock and campaigns against, instead of annual averages. The model also surfaced effects nobody was planning for, such as a sales rise in one region during hurricane season, now treated as a predictable seasonal peak. Accuracy figures are withheld under the client agreement.
Engineers buying load cells and controllers needed to find the right part by application, request quotes and pull certificates, and the support team was answering the same technical questions by hand from a library of 900+ articles.
A nopCommerce B2B platform with navigation by weighing application, configurable product pages and dynamic quote requests; SQL synchronisation feeding a certificate generator; and an AI knowledge base with retrieval-augmented search across the article library.
Customers self-serve on part selection, quotes and certificates, and support answers the questions that remain from the same 900+ articles in half the time.
A distributor serving a national carrier’s retail network sold 1,100+ SKUs through an ageing store that did not talk to its ERP, so inventory, orders and payments were reconciled by hand.
A rebuilt nopCommerce store with tiered pricing and a mobile-first checkout, two-way NetSuite integration for products, inventory and orders, and scheduled ACH payment processing through NACHA files with the client’s bank.
Retailers order and pay online, inventory and orders stay in step with NetSuite, and finance stopped re-keying payments.
Agents at a coastal brokerage asked the same questions of an outdated knowledge base, and the answers lived in documents nobody could find quickly.
A retrieval-augmented assistant on AWS Bedrock with Elasticsearch as the search layer, integrated into the brokerage’s WordPress intranet. It answers from company resources and the web, with content kept in sync by the admin team.
Agents get answers from company material in seconds instead of asking around, and the admin team controls what the assistant knows.
A new ride-hailing service needed a backend that could match drivers to riders in seconds and stay consistent while many bookings change at once.
A microservices backend on AWS ECS with real-time updates over SignalR, driver matching protected by Redis distributed locks, and booking and payout logic in PostgreSQL, serving the customer, driver and administrator apps.
The platform is in progress with the client: backend and APIs are built and documented for the Flutter mobile team, with the service preparing for launch.
A global IT distributor offered thousands of vendor certification courses through Cornerstone OnDemand, but finding the right class was slow and the LMS could not be replaced. We built a portal layer with auto-complete search, IP-based localisation and a responsive interface, with enrolment still handled in Cornerstone.
An industrial cord supplier moving online needed customers to find cords by gauge, length, rating and jacket. We built a nopCommerce store with a custom configurator and Apache Solr attribute filtering.
A manufacturer of emergency access systems sells to fire departments, law enforcement and property owners, each with its own rules. We built a Kentico commerce site with Salesforce integration, a customer portal, an order-approval workflow, dynamic PDF generation and a fire-department lookup.
A role-based portal with a searchable library of thousands of parts and dynamic order forms for custom covers and lids, replacing orders by phone, mail and fax.
Custom API integrations deliver parcel tracking, drop-box lookup, rate calculation and account login on a site the carrier’s own teams manage.
Programmes consolidated onto one site, with a custom API integration to the PowerCampus student information system and enrolment-focused search.
The stack our team has used in production over 18+ years, with AI and automation now at the centre of most new engagements. Dark tiles mark our core, where we are strongest.
We choose the stack for the problem, and we are candid when something outside this list is the better fit.
Four stages, each ending with something concrete in your hands. You can stop after any of them.
We learn your business, your data and your constraints, then map options with honest estimates for each.
Architecture, data model and interface decisions made in the open and reviewed with you before production code starts.
Working software lands in a live environment every week, so you steer with real screens rather than status reports.
Monitoring, support and steady improvement after launch, on a retainer sized to how much the system changes.
Smaller than the big consultancies, more senior than the freelance marketplaces.
The engineers who scope the work are the engineers who build it. No hand-off to a junior bench after the sale.
We have shipped AI systems that process real documents and real orders every day. Prototypes are a step, never the deliverable.
Commerce platforms, enterprise integrations and data systems on AWS and Azure, built by engineers with 18+ years behind them. We fit the systems you already run instead of replacing them.
KodeLogiK has served American companies since 2022, and its founder for nearly two decades before that: overlapping hours with both US coasts, contracts and invoicing where you need them, and a track record with distributed client teams.
Two simple ways to engage. Many clients start with a project and move to a dedicated resource once the system is live.
We scope the work with you, quote it project by project, and deliver against that scope with weekly releases. Changes are agreed in writing before they affect cost or timeline.
A senior engineer, or a small team, embedded with yours on a monthly basis: working in your tools, your sprints and your hours, with continuity from month to month.
Hiten has spent 18+ years designing and delivering software for US companies, most recently as Technical Architect at established software firms, with a focus on applied AI, machine learning and e-commerce platforms. His work spans B2B commerce systems integrated with ERPs, retrieval-augmented knowledge bases and agentic systems built on the latest Claude and OpenAI models, and generative features built into live retail stores. He founded KodeLogiK in Pune in 2022 so that clients work directly with the architect who designs their system and remains accountable for it through delivery and beyond.
How we approach the problems we are asked about most, written by the engineers who do the work.
Most companies do not need a new AI platform. They need their existing .NET systems to be callable by a model, safely. Here is the architecture we use.
Annual averages hide the patterns that run an online store. A forecasting model built on your own order history can see them, and it does not need a new platform.
Connecting a store to an ERP is less about the API than about deciding who owns what. These are the rules we build to, and the failures they prevent.
The questions we hear most on a first call.
A short call is enough to tell whether we are the right fit. Send a few lines about the project, or write to Hiten directly; replies come from him, usually within one working day.
