Independent AI Solutions Architect
AI-assisted product architecture and development using React, FastAPI and PostgreSQL access controls. Translate requirements into product specifications, source-backed outputs and explicit review paths.
BANGKOK, THAILAND / THEERAWEE RANGRAM
Before I built systems,
I imagined spaces for people.
Boom · AI Engineer & AI Solutions Architect
I design and build AI experiences that make complex work feel clear, approachable, and inviting to use.
01 / THE FIRST BLUEPRINT
Before AI · the urge to design
As a child, I wanted to be an architect—to design spaces people could step into and inhabit.
That early pull toward design stayed with me when I chose to study software engineering. Frontend became my creative space—a place where code took visible form, where people could touch, test, and make it their own.
I care about the character of what I build. The proportion of an interface, the rhythm of its interactions, and the clarity of its next step all shape how it feels to use. Software simply gave that same architectural impulse a different medium.
Designing for a screen was only the start; real software has to hold its weight in daily use.What happens when people begin to rely on what you build?02 / FOUNDATIONS
Before AI · frontend foundations
Frontend that has to work as part of a system people depend on.
Enterprise frontend work at Allianz Technology gave my interest in design a practical discipline. A single screen is never an island; it lives inside a larger application with strict requirements, shared standards, and real users relying on it.
Collaborating across teams in Munich and Bangkok taught me to look beyond surface polish. The quiet details a user never consciously notices are often the ones an engineer has to get right.
Allianz TechnologyAngular and enterprise frontend development, working with stakeholders and engineering teams in Munich and Bangkok.
Reliable components are only half the craft; they still have to fit into a wider architecture.How do you build when your screen is just one piece of a much larger effort?03 / CONNECTIONS
Before AI · working across a larger system
What I built on screen had to connect with other people’s work.
Corporate banking work at Accenture sat right where business intent meets hands-on engineering. My role lived inside a shared delivery effort—listening closely to what stakeholders actually needed, translating those requirements into frontend architecture, and building alongside engineers across locations.
When squads work across different places, shared clarity matters just as much as clean code. It trained me to see the whole structure first, so the piece I owned would fit naturally into everyone else’s.
AccentureCorporate-banking frontend delivery, translating stakeholder requirements and collaborating with distributed engineering squads.
In 2024, while I was still working at ttb, I began developing my AI capabilities.What if intelligence could become part of the workflow?04 / THE POSSIBILITY BECOMES PERSONAL
Beginning AI development while still working at ttb
I began developing my AI capabilities while I was still at ttb. What I believed still needed to take shape as something people could test in real work.
Some people around me were not yet convinced, including people working in the same field. I wanted to give them something they could actually try. My belief alone could only take the idea so far.
I started with the work in front of me: bringing GPT into Jira, Confluence and Bitbucket workflows, with people reviewing the outputs. The possibility I had imagined began to present real problems I had to work through.
A model’s answer was a starting point. I still needed to understand where it came from, what was missing, and where a person should stop and inspect it. Working through those details showed me the distance between an answer and a tool someone could trust enough to use.
ttb sparkSenior software engineering in banking, including frontend delivery. Began developing AI capabilities in 2024 while at ttb, with GPT-assisted Jira, Confluence and Bitbucket workflows and human review.
I wanted to give them something to try before asking them to believe.Beneath the visible progress lay real friction—edge cases, false starts, and moments when the path forward was unclear.What did I do when I could not yet see the next step?
05 / ANOTHER SIDE OF THE BEGINNING
A closer look at the AI work that began at ttb
As I began learning and developing AI, some people around me did not yet see the same possibility.
Even peers in the same discipline were unconvinced—and fairly so, while the idea existed mostly in theory. To make it tangible, I started building on my own, working through each obstacle one piece at a time.
Every skill and workflow I create traces back to those hands-on encounters. I test an approach against a real problem, examine where it breaks, look for sharper methods, and return to refine the next version. Week after week, that steady cadence became HolyMind.
Every difficult part asked something of me. Every return to the work left something I could build on.The problems I encountered gradually became lessons I could carry into the next piece of work.How does a lesson become a skill with a character of its own?
06 / EXPERIENCE, DISTILLED
Lessons gathered through developing AI
I let my experience developing AI settle, then ask what is worth keeping.
Once a lesson holds up in practice, I look outward. I study research, engineering patterns, and tools from around the world, selecting and combining the strongest methods I can find for the problem in front of me.
Through repeated testing and trimming, each skill earns its own purpose, method, and character. Connected into workflows, those capabilities give new work a stronger foundation from the start.
That is the moat I am building: not a secret shortcut, but lived experience, deliberate curation, and revision after revision until a capability proves its worth in real use.
A visual metaphor for learning and synthesis. The scenes are symbolic, not documentary recreations.
Follow these capabilities into the work All that depth still needs a way for someone to reach it.How do you make all that complexity feel human?07 / WHERE IT COMES TOGETHER
Today · design, software and AI together
I want people to feel the possibility, without carrying all the complexity.
Working with AI brings my frontend instincts and software engineering background together in a wider creative space. I can shape both the underlying machinery of a tool and the way someone enters, understands, and works with it.
The interface should absorb the heavy complexity behind the scenes while keeping the next step clear, the interaction inviting, and the person in control. Reasoning, sources, and moments for human review remain within reach whenever they are needed.
Thoughtful design opens the door, but care in every detail of use is what makes people want to return. To me, that is the heart of AI experiences.
Independent AI Solutions ArchitectAI-assisted product architecture and development using React, FastAPI and PostgreSQL access controls. Translate requirements into product specifications, source-backed outputs and explicit review paths.
A clear starting point, meaningful feedback, and a next step you can understand.
Applications, models, and tools connected around the task.
Explicit permissions, inspectable sources, and human review paths.
An illustration of my design approach, not a live system diagram.
08 / OPEN TO EXPLORATION
Current public work
HolyMind interactive portfolio
Step into a browser-based demonstration of how human intent, workflow boundaries, and interface design work together. Explore three illustrative scenarios, trace fictional source-to-output examples, and carry your selections into a planning brief.
This is a portfolio demonstration, not a live agent service or a claimed client result. Its browser behavior is open to inspect here; client source code, private records, and confidential documents are never published.
FOLLOW A CAPABILITY INTO THE WORK
A repetitive request still needs an accountable owner and a clear review boundary.
Make the draft, missing information and next decision visible together.
Explore the repetitive-work scenario, inspect the example input, then review the proposed output.
A confident answer is difficult to assess when its supporting information is hidden.
Keep claims connected to approved excerpts and leave unsupported facts unanswered.
Open the scattered-knowledge scenario and follow a draft claim back to its fictional source.
Exploring a collection should lead somewhere useful when a capability catches your attention.
Keep the visitor’s scenario and chosen skill connected to the next conversation.
Choose a scenario, explore a HolyMind skill, then carry that skill into the starting brief.
09 / THE PROFESSIONAL VIEW
Selected experience, most recent first. Detailed CV and project context are available in a recruiter conversation.
HolyMind
34 skills shaped through practice.
AI-assisted product architecture and development using React, FastAPI and PostgreSQL access controls. Translate requirements into product specifications, source-backed outputs and explicit review paths.
Senior software engineering in banking, including frontend delivery. Began developing AI capabilities in 2024 while at ttb, with GPT-assisted Jira, Confluence and Bitbucket workflows and human review.
Corporate-banking frontend delivery, translating stakeholder requirements and collaborating with distributed engineering squads.
Angular and enterprise frontend development, working with stakeholders and engineering teams in Munich and Bangkok.
Translate business requirements into system boundaries, integrations, access controls and a delivery plan.
Build AI-assisted workflows that connect models, tools and existing applications, with evaluation and human review.
Bring customer-facing requirements, hands-on software delivery and international stakeholder coordination into one implementation conversation.
Identify a useful first use case, define the workflow and governance, and agree how its results will be evaluated.
These are relevant roles to discuss, not a list of job titles I have held. Scope, availability and commercial terms are agreed in conversation.
Bangkok, Thailand. My enterprise background includes collaboration with teams in Bangkok and Munich.
No. My background is enterprise software delivery; I began developing AI capabilities in 2024 while at ttb. My current focus is applied GenAI, AI solutions architecture and agentic automation—not a claim of nine years of foundation-model research.
Use LinkedIn with the role, team, location and problem you need solved. We can discuss fit and the appropriate supporting project context.
10 / THE NEXT CHAPTER
I still build places for people. The medium has changed.
Bring the problem your team is trying to solve. Let’s explore an AI experience people can understand, trust, and want to use.
Return to the first line