Business applications, payment operations and applied AI: professional experience behind the work Hamsa offers today.
These examples describe Hamsa Jama’s contributions in previous professional roles, including work within wider teams. Figures are approximate, reported in his professional record, and reflect the original project context.
The portfolio spans business applications and workflow automation; payments, subscriptions, integrations and finance operations; and applied AI, document processing and prediction-related applications.
01 / Cazoo
Give finance its time back.
Senior Software Engineer · 2021–2024
120+ → ~20 minweekly revenue-reporting process
The problem
A weekly revenue-reporting process required more than two hours of manual work.
Hamsa’s contribution
Designed and delivered an AWS serverless revenue-reporting pipeline, working with finance, product and QA through requirements, testing and production delivery.
The reported outcome
The weekly process fell from 120+ minutes to approximately 20 minutes.
Why this matters for a smaller business
For a smaller business, the useful starting point is often one recurring report: where its data comes from, who checks it, and what still needs a person’s judgement.
Payment reconciliation was a multi-day finance process and a persistent operational bottleneck.
Hamsa’s contribution
Built a serverless payment reconciliation service. The wider role included finance automation and NetSuite ERP integrations in a regulated, audited environment.
The reported outcome
The reconciliation process was compressed from multiple days into minutes.
Why this matters for a smaller business
The same kind of integration work can help a smaller team bring payment and finance records together. The right solution depends on its systems, volumes and exceptions.
~£60k / monthreduction in payment processing costs
The problem
A legacy payment provider needed replacing, alongside work to bring subscription capabilities in-house.
Hamsa’s contribution
Led the migration to Stripe and designed and built an in-house subscription platform to replace a third-party platform. Also delivered Apple Pay integrations and purchase-flow instrumentation.
The reported outcome
The payment migration reduced processing costs by approximately £60,000 per month. The in-house subscription platform also gave internal teams direct control over subscription logic and roadmap.
Why this matters for a smaller business
An SMB may have a much smaller bill, but the questions are familiar: what are you paying for, where does the integration create work, and which changes are worth the cost?
Document processing took a week before the resulting information could support downstream analytics.
Hamsa’s contribution
Designed and built named-entity extraction services using Python and spaCy, alongside backend services and data pipelines.
The reported outcome
Document-processing time was reduced from one week to one hour.
Why this matters for a smaller business
For an SMB with repetitive document work, this is a useful pattern to assess: extract the fields that matter, validate them, and give exceptions a clear review path.
Hamsa Jama contributed to these projects while consulting at Tata Consultancy Services. The organisations were end clients in that context; these descriptions do not imply contracts with or endorsements of Hamsa Software Ltd.
PwC · Internal holiday-booking application
Reached production. Contributed to an internal application for booking holidays.
PwC · NLP and machine-learning document redaction
Reached production. Contributed to an internal document-redaction tool using natural language processing and machine learning.
Apple · Forecasting and prediction work
Reached production. Contributed to forecasting/prediction work.
SELECTED TCS CONSULTING PROJECTS
Explore before committing.
Hamsa Jama contributed to these projects while consulting at Tata Consultancy Services. The organisations were end clients in that context; these descriptions do not imply contracts with or endorsements of Hamsa Software Ltd.
Microsoft · Root-cause-analysis application
Proof of concept. Contributed to a proof-of-concept root-cause-analysis application.
A Dubai-based government office · OCR for scanned Arabic documents
Proof of concept. Contributed to a proof-of-concept OCR engine investigating text extraction from scanned Arabic documents.
PROFESSIONAL INTEGRATION EXPERIENCE
Across the payment and finance journey.
NetSuite · Zuora · Stripe · Salesforce
Experience relevant to the handover
Hamsa has worked with NetSuite, Zuora, Stripe and Salesforce integrations. Together, these platforms cover finance records, subscriptions, payments and customer information—the places a payment workflow often has to join up.
What that means for an engagement
We can discuss the business meaning of the records as well as the APIs: identifiers, amounts, fees, subscription events, exceptions and who owns the next decision. The scope is based on your systems and access, rather than an assumption that every configuration works the same way.
The boundary of this claim
This is professional integration experience. It does not imply an official platform partnership or that a ready-made AI Institution connector exists for each system. A customer pilot needs its own implementation and acceptance checks.
Question↓Current, permitted documents↓Cited answer or human handoff
The business question
Can a team find an answer in its policies and guides, see where it came from, and recognise when the available evidence is insufficient?
What the prototype demonstrates
The Receipts RAG project combines document search with a bounded answer-checking loop. It withholds superseded documents, demonstrates role-based filtering, and returns source citations or a handoff when it cannot support an answer.
What has been checked
The repository includes synthetic support and employee-handbook examples, evaluation cases and a recorded offline demonstration. The offline model is a deterministic test adapter; its results check the plumbing, not real-world AI answer quality.
Where it is today
This is a learning and evaluation prototype, not a production customer-support service. Its role filtering is a demonstration, not a complete authorisation system. A business deployment would need its own access controls, document review and evaluation on representative questions.
