AI Ecosystem Architect | I build LLM agents, autonomous trading systems & self-improving ML platforms that work in production — not just in demos.

I help startups and product teams ship complete, production-grade
AI systems — autonomous trading bots, LLM agents, self-improving
ML platforms, and enterprise analytics tools — that run reliably
when real users and real data hit them.

With 20+ years of engineering experience and a Physics degree
from Bandung Institute of Technology (ITB), I architect full
ecosystems from the ground up: backend, AI layer, infrastructure,
and frontend. Solo, end-to-end, with full test coverage.

My core engineering principle: AI explains, math computes. Every
number is deterministic and verifiable. No hallucinated
statistics. Ever.

WHAT I'VE SHIPPED:

▸ Agent Scientist Studio — Self-improving ML platform with ACE
(Analyze-Curate-Enrich) architecture. Messy CSV in, deployed
ML model out — no code required. AutoML, SHAP explainability,
one-click REST API deployment. Backed by a published
technical white paper.

▸ Antigravity Trading Suite — Complete autonomous Forex + Crypto
trading ecosystem: real-time AI trading bots, Go-based
high-concurrency data pipeline, Guardian monitoring with
real-time alerting, live Next.js control panel. Operational
24/7 with zero silent failures.

▸ Aura Data Platform (ACE) — Enterprise AI analytics platform:
~16,500 lines of code, 914 automated tests, 49 API endpoints.
Natural language queries, 5 forecasting methods with confidence
intervals, self-learning knowledge base via PostgreSQL 17 +
pgvector. Live at aura-aitech.com.

▸ HR AI Demo — RAG-powered conversational AI automating 60% of
routine HR inquiries. Grounded answers from company knowledge
base — no hallucinations.

▸ Travel Dream Simulator — AI-powered personalized trip planning
with hidden-gem algorithm. 45% improvement in trip choice
satisfaction in user testing.

HOW I WORK:
→ Design-first: technical architecture doc before a single
line of code — no surprises mid-project
→ Fallback-first: every LLM call has a rule-based fallback —
your product never dies because an API went down
→ Test-driven: 900+ automated tests across my portfolio —
refactors are safe, bugs surface early
→ Full-stack ownership: FastAPI · Go · PostgreSQL/pgvector ·
LangChain · Next.js · Linux · Docker
→ One point of contact — you don't juggle three contractors

I start every engagement with a small, well-scoped paid
milestone so you can verify quality before any larger
commitment. Low risk for you.

Portfolio & live products: aura-aitech.com

Work Terms

AVAILABILITY:
Available more than 30 hours/week. Response time: 0–4 hours
(GMT+7, Bandung, Indonesia). Overlap available with US
mornings, EU afternoons, and AU/NZ business hours.

HOURLY RATE:
$55/hr for ongoing hourly engagements.
Fixed-price available for well-scoped projects.

HOW I START EVERY PROJECT:
1. Discovery call (30 min) — understand your goals and
technical context
2. Technical design document — architecture, API design,
database schema, AI strategy — before any code
3. Milestone-based delivery — you approve each phase
before I move to the next
4. Testing & handover — documented code, test suite,
and deployment guide included

PAYMENT TERMS:
- Hourly: weekly billing via Guru SafePay
- Fixed-price: 50% upfront, 50% on delivery
- First milestone: happy to start small so you can
verify quality before larger commitment

COMMUNICATION:
- Primary: Guru messages + video calls (Zoom/Meet)
- Daily async updates on active projects
- Fluent English + Native Indonesian (Bahasa)

WHAT I DO NOT TAKE ON:
- Projects without a clear scope or brief
- "Build me something like ChatGPT" without specifics
- Jobs requiring misrepresentation of AI capabilities