I build AI systems.
And the teams
behind them.

I began in the automotive space, designing and delivering 20+ RPA bots for a global brand. The work connected systems that were never designed to work together—Citrix, Mainframe, Oracle EBS, Microsoft Dynamics, email, and complex desktop workflows.

At R Systems, I led a greenfield automation program for a major Japanese automotive organisation. I was deployed onsite as Lead Engineer and took ownership from process discovery and architecture through production environment setup and delivery.

The flagship import-export automation dealt with genuine operational complexity: government notifications, customs, tax, INCOTERMS, invoices, Free Trade Agreements, and highly variable source documents. The automation handled document extraction and the variations involved in import-export processing.

I also automated procurement and invoicing—from consolidating RFQs and identifying anomalies to plant assignment, vendor reconciliation, Oracle EBS uploads, IRN generation, and indexing. These workflows covered RFQ matching, anomaly detection, purchase-request checks, and invoice reporting.

At Zensar, I worked with a leading U.S. modem and communications-device manufacturer to automate core finance operations. The mandate covered accounts payable, accounts receivable, debit notes, and chargebacks—processes where accuracy and auditability are as important as speed.

I delivered four end-to-end automations across email, Oracle EBS, Salesforce, and internal business applications. The implementation used Automation Anywhere and IQ Bots to process documents and move information between applications.

The automations scaled to more than 100 vendors, eight geographies, and eight languages. This extended the same four finance workflows across a multilingual vendor base.

Earlier and later work deepened that finance lens through reconciliations spanning Excel, SAP, APIs, and databases, plus automated invoice generation and vendor matching. The reconciliation bot brought records from these sources into a single automated matching process.

With a Fortune 500 EdTech partner, we moved beyond single-purpose AI features to build a coordinated set of learning agents. The engagement covered both educational content and administrative workflows.

Using AutoSage v2 and its API layer, we created agents for generating learning material—up to 100 pages at a time—question papers, presentations, and exam evaluation. The work used early multi-agent patterns to make complex content workflows more deliberate and repeatable.

The platform also supported faculty invoicing and field-sales training, extending the value from classroom content into the operating machinery around education. These agents addressed separate teaching, administration, and sales use cases.

I am also a co-inventor on four jointly filed patents in the EdTech space. The patent applications relate to the EdTech work; they are filed applications, not a claim of granted patents.

For a Berlin-based venture capital fund, I led the development of AgentOS: an interoperating autonomous system designed to support the full lifecycle of a company. It brought company formation, HR, finance, sales, and marketing into one agentic operating layer.

The engagement also included a company-brain initiative, giving teams a more useful way to access and act on organisational knowledge. This was a separate project alongside AgentOS, startup screening, and outreach.

We also built startup screening and scoring capabilities based on a proprietary evaluation mechanism, alongside CRM and AI-powered outreach. The screener implemented the fund’s proprietary scoring mechanism; the CRM supported its outreach workflow.

Alongside this work, I have helped early products find their footing—from a legal-data chatbot for state and county websites to an Australian AI-powered PMS centralisation app and an intelligence layer for an AdTech partner. The legal-data and PMS projects were smaller engagements preceding the larger platform work.

My work also includes a voice AI product for outbound calling. At Autonomix, I architected FirstLine, a communications platform spanning inbound and outbound engagement.

In the HR benefits space, I helped shape a joint venture powered by AutoSage v4. The joint venture used AutoSage v4 as its AI foundation.

I have also worked on a revenue-cycle management product focused on eligibility and denial management. The scope covered eligibility checks and denial-management workflows.

Across joint ventures, my responsibilities include shared architecture, engineering direction, infrastructure costs, and compliance requirements.