AI IMPLEMENTATION / PRODUCT SYSTEMS

Between people and machines, I make the handoff work.

I turn ambiguous workflows and messy data into bounded prototypes, decision systems, and implementation plans.

J. Kajeesan signature

SELECTED WORK

OPENHEALTHATLAS

From fragmented records to a bounded, auditable decision workflow.

MY ROLE
PRODUCT / SYSTEMS / PROTOTYPING
Sanitized case study
  1. BUILT

    A governed evidence workflow using fictional records.

  2. DECIDED

    Separate deterministic records and calculations from external AI interpretation.

  3. PROVED

    A substantial reference implementation with fictional acceptance evidence for readiness, provenance, evidence replay and bounded tool use.

Read project notes

OpenHealthAtlas is a governed evidence system that turns fragmented health records into deterministic findings an external AI can investigate without becoming the source of truth.

It demonstrates product architecture, validation, provenance and bounded AI tool use with fictional data.

A three-stage paper diagram showing health-data gathering, deterministic evidence validation, and a bounded Hermes analysis report.
Synthetic OpenHealthAtlas workflow demonstration

SELECTED WORK

CAPITO CAPITAL

In one self-reported applied comparison, the prototype produced structured output in under two hours; the prior process was described as taking approximately one month.

MY ROLE
AI WORKFLOW / RESEARCH SYSTEM / PROTOTYPING
Applied prototype · independent work
  1. BUILT

    A structured investor-research workflow.

  2. DECIDED

    Define the reviewable output before automating search.

  3. PROVED

    One self-reported prototype run produced structured output in under two hours; it was not independently timed or repeated.

Read project notes

A student worker had previously spent approximately one month manually finding and structuring potential-investor information. I investigated the desired output and directed an AI-assisted data-extraction prototype.

This is a qualified, self-reported applied comparisonnot a claim of recurring production performance, verified savings, paid commercial validation or investment outcomes.

A paper-based investor research workflow with source documents connected to a structured review sheet.
Conceptual view of the structured research workflow.

SELECTED WORK

CPNB CATHETER THESIS

From 25 concepts to one experimental catheter-anchor prototype and a bounded test approach.

MY ROLE
PRODUCT DEVELOPMENT / PROTOTYPING / TEST DESIGN
Joint MSc thesis with Tobias Sebastian Bernhardt
  1. BUILT

    Prototypes, fixtures and a dedicated test rig.

  2. DECIDED

    Narrow concepts using explicit criteria and observed failures.

  3. PROVED

    A 6.5 N maximum in one documented beef-tissue comparison conducted with physicians, versus 1.5 N for the straight control in that setup.

Exploratory physical test · not clinically validated

Read project notes

With Tobias Sebastian Bernhardt, I moved from clinical research and 25 concepts through structured selection, prototyping, fixtures and test-rig development to one anchor concept.

The 6.5 N value is the maximum recorded result for the anchor in one documented beef-tissue comparison conducted with physicians, versus 1.5 N for the straight control in that setup. It is not an average or a claim of efficacy, safety, repeatability, production performance or clinical validation.

Catheter-anchor concepts, a blue-and-white prototype, and technical sketches showing exploded, cross-sectional and alternative anchor geometries.
Continuous peripheral nerve block (CPNB) catheterExploratory thesis evidence

04 / WORK EXPERIENCE

Across products, people, data and AI implementation.

A career shaped by learning unfamiliar domains, making boundaries visible, and turning complex work into decisions teams can explain and act on.

  1. OCT DEC 2019

    Product Development Intern

    Novo Nordisk · Denmark

    Developed and iterated a physical measurement concept for insulin-pen clearance and alignment work.

  2. SEP 2021 AUG 2022

    Graduate Teaching Assistant

    Technical University of Denmark

    Helped multidisciplinary teams turn problem framing and user needs into testable technical and business propositions.

  3. JAN DEC 2024

    Test Consultant

    Netcompany · Copenhagen

    Created and executed repeatable migration tests for confidential customer data in a 50+ contributor programme.

  4. JAN SEP 2025

    IT Consultant

    VOLI · Copenhagen

    Supported Maconomy-based ERP implementation and customer-facing SaaS delivery.

  5. OCT 2025 PRESENT

    Independent AI Product & Systems Development

    Independent · Copenhagen

    Directed AI-assisted systems through product definition, evidence boundaries, prototyping and acceptance criteria.