Our Solutions
Three engagements with defined scope and measurable output
We do not offer open-ended consulting arrangements. Each engagement type has a documented structure, a stated duration, and a set of deliverables that are agreed before work begins.
← Back to HomeMethodology
How every engagement is structured
Every engagement — regardless of type — follows the same structural pattern. It begins with a written scope document that specifies what will be delivered, how it will be evaluated, and what is outside scope. No work starts before that document is agreed.
During the engagement, we work alongside your team rather than separately from them. Design decisions are explained as they are made. Questions from your engineers are answered directly. We treat knowledge transfer not as a phase at the end of the engagement but as something that happens continuously throughout.
At the close of the engagement, the deliverables include the technical work itself, written operational documentation, and — where applicable — the evaluation methodology in reproducible form. You should be able to assess the quality of what was delivered without relying on our word for it.
Written scope first
Deliverables, timeline, and success criteria agreed in writing before any work begins.
Collaborative throughout
Your engineers are present and involved, not waiting for a handover at the end.
Measurable outcomes
Evaluation methodology defined before work begins and delivered as output.
Documented decisions
Operational documentation covering design choices and known limitations.
Engagement 01
Production Engineering for ML Systems
A specialist engagement for Malaysian firms whose ML prototype has produced demonstrably useful outputs in controlled conditions but has not yet been built to the standard required for sustained production use. We examine the prototype, identify the failure modes that production will surface, and work alongside the internal team to build the monitoring, retraining, and evaluation infrastructure the system needs to operate over time.
- Prototype audit: input distribution, output quality, failure modes
- Monitoring and alerting infrastructure built alongside your team
- Retraining pipeline with evaluation gates before promotion
- Operational documentation and runbook
Process steps
Prototype review — examine the existing system, data pipeline, and evaluation methodology (if any). Document failure modes and production gaps.
Scope agreement — agree in writing on which gaps will be addressed in this engagement and how success will be measured.
Engineering phase — build monitoring, retraining, and evaluation infrastructure alongside your team, explaining decisions throughout.
Documentation and handover — deliver operational documentation, evaluation methodology, and a structured handover session.
MYR 2,250
Fixed fee · 6–10 weeks typical
Engagement 02
Retrieval and Search Re-engineering
A focused engagement for organisations whose internal search and retrieval systems have grown unsatisfactory. We assess the current implementation and the document corpus, establish a working definition of what good retrieval means for your specific information needs, measure the current system against that definition, then re-engineer the retrieval pipeline using current methods — typically a combination of careful chunking strategy, embedding model selection, and post-retrieval ranking.
- Corpus analysis and retrieval quality baseline measurement
- Re-engineering: chunking, embedding selection, post-retrieval ranking
- Explicit, reproducible measurement methodology as deliverable
- One round of post-launch tuning included
Process steps
Assessment — examine the corpus structure, current pipeline, and failure cases. Establish what good retrieval would look like for your actual information needs.
Baseline measurement — measure the current system against the agreed definition of good retrieval. Record results reproducibly.
Re-engineering — implement chunking, embedding, and ranking improvements. Re-measure at each significant change.
Post-launch tuning — one follow-up round after real-user traffic, addressing gaps that only appear under production usage patterns.
MYR 1,180
Fixed fee · includes post-launch tuning
Engagement 03
Engineering Mentorship for Internal Teams
A standing arrangement for in-house engineering and data teams who are taking on responsibility for AI work but lack a senior counterpart with AI engineering experience to consult. We provide a weekly structured review of the team's current work, ad-hoc availability for specific technical questions during the week, and quarterly written assessments of the team's developing judgment and practice.
- Weekly structured code and decision review session (60 minutes)
- Ad-hoc availability for technical questions during the week
- Quarterly written review of team progress and developing practice
- Initial six-month term, renewable
How the weekly sessions work
Team shares current work — active code, design decisions under consideration, or problems encountered that week.
Structured review — we examine the work together, ask the team to articulate their reasoning, and offer alternative framings where useful.
Technical questions — specific questions raised during the week are addressed in the session or via written response the same day.
Quarterly written assessment — a summary of the team's progress, recurring strengths, and areas where judgment is still developing.
MYR 590 /month
Six-month initial term · renewable
Decision Guide
Which engagement fits your situation
The three engagements address different problems. If you are unsure which one applies, describe your situation in the contact form and we will advise.
| Your situation | Production Engineering | Retrieval Re-engineering | Team Mentorship |
|---|---|---|---|
| ML prototype exists, needs production hardening | ✓ Primary fit | — | — |
| Internal search quality is below expectations | — | ✓ Primary fit | — |
| Team taking on AI work without senior AI counterpart | — | — | ✓ Primary fit |
| Need monitoring and retraining infrastructure | ✓ Included | — | — |
| Multilingual corpus (Malay / English / Chinese) | Relevant | ✓ Specialist area | — |
| Fixed-fee, defined scope preferred | ✓ Fixed fee | ✓ Fixed fee | Monthly |
| Ongoing structured support needed | — | — | ✓ Standing arrangement |
Standards
Applied across all engagements
Confidentiality and data handling
A signed confidentiality agreement is in place before any client data or system information is shared with us. Client data is not retained after engagement close.
PDPA compliance awareness
We work within Malaysia's Personal Data Protection Act framework and factor data handling requirements into system design recommendations.
Version-controlled deliverables
All code and configuration produced during an engagement is delivered in version-controlled form with commit history intact.
Clear communication norms
Technical questions from your team receive same-day responses during the engagement period. Progress updates are provided at agreed intervals.
Operational documentation standard
Documentation covers: what was built, why each significant decision was made, known limitations, and what to do when specific failure conditions occur.
IP stays with you
All work product produced during the engagement belongs to your organisation. We retain no rights to the systems or documentation produced.
Pricing
Fixed fees, invoiced in Malaysian Ringgit
Engagement 01
Production Engineering
MYR 2,250
Fixed fee · 6–10 weeks
- Prototype audit and gap analysis
- Monitoring & alerting setup
- Retraining pipeline with evaluation gates
- Operational documentation and runbook
- Knowledge transfer throughout
Engagement 02
Retrieval Re-engineering
MYR 1,180
Fixed fee · includes post-launch tuning
- Corpus and pipeline assessment
- Baseline retrieval quality measurement
- Chunking, embedding, and ranking improvements
- Reproducible evaluation methodology
- One post-launch tuning round
Engagement 03
Team Mentorship
MYR 590 /mo
6-month initial term · renewable
- Weekly structured review session (60 min)
- Ad-hoc technical question availability
- Quarterly written progress assessment
- Session built around actual team work
- Renewable at end of each term