Rising Seer — AI-Native Fractional CTO
Fractional CTO for an AI-native astrology platform — leading infrastructure, tooling, and delivery toward launch.
risingseer.comTechnologies Used
Key Features
What Rising Seer is
Rising Seer is a mobile-first astrology platform combining polished UX with deep astrological reasoning, running on a multi-agent AI pipeline. I joined as Fractional CTO heading into the launch window — priority engagement on a monthly retainer, daily standup cadence, end-to-end responsibility for technical direction.
Stack: Expo / React Native mobile, Python / FastAPI backend with a Flask microservice cluster behind a gateway, multi-provider LLM pipeline (OpenAI + Gemini) with hybrid retrieval (Pinecone + BM25), Firebase + GCP, Terraform IaC.
My role — the four pillars
Leadership engagements at AI-Native Works run on four pillars. Rising Seer is where that framework shows up in practice.
Roadmap & Strategy
Technical direction aligned with business outcomes. What to build, in what order, and why.
- End-to-end codebase technical review — architecture, AI pipeline, data model, security, CI/CD gaps, scalability ceilings. The resulting risk register and workload assessment anchored subsequent decisions.
- Feature-scoping for two major initiatives (context knowledge graph, RAG retrieval overhaul) — breaking design docs into projects, milestones, and tickets the team could execute against.
- Team composition and sizing strategy for an AI-accelerated workflow — deliberately small and leverage-forward rather than headcount-forward.
Architecture & Tooling
System design, technology selection, AI integration strategy. The decisions that determine whether a product scales or stalls.
- Staging environment stood up from scratch — separate Cloud Run services, Firestore instance, Pinecone namespace.
- Tag-based production deploy pipeline with GitHub Environment approval gates, paired with staging-on-merge for fast iteration.
- PostHog feature-flag plumbing enabling a flagged-off-on-main pattern — no more branch handoffs between engineers.
- Single-source-of-truth version architecture resolving mobile OTA vs. native build ambiguity.
Developer Enablement
Introducing AI-native practices to the existing team. Not replacing engineers — making them dramatically more effective.
- GitHub Issues to Linear migration — 200+ issues moved with dependencies intact, plus a custom Linear integration for AI coding agents to reduce token usage.
- Onboarding and technical evaluation framework for the incoming full-stack engineer.
- Daily standup and PR-review cadence — the operational heartbeat as velocity picks up.
Output Validation
Quality gates, architecture review, production readiness. Ensuring what ships is production-grade.
- Structural PR review on the hybrid-retrieval RAG rollout — architectural defects caught before main.
- PR review across the team as standing practice, holding the code-quality bar under accelerating AI-assisted velocity.
- Closing CI gating gaps: pytest failure suppression, Firestore composite-index parity between test and prod environments.
Scope
- Monorepo spanning mobile, Python backends, multi-agent AI pipeline, and Terraform infrastructure
- Multi-provider LLM orchestration with hybrid retrieval, hedged calls, and custom context assembly
- Fractional engagement on monthly retainer, priority availability, daily standup cadence
- Small team: founders, one full-stack engineer, one QA analyst, AI-accelerated delivery model
Recent wins and ongoing
- Team composition reset — surfaced and resolved contractor misalignment ahead of launch; restructured around a smaller, higher-leverage team
- Staging environment and tag-based production pipeline shipped
- Terraform coverage audit completed, infrastructure gaps closed
- Single-source-of-truth version architecture in review
- Linear workspace migration and full backlog restructure
- Codebase technical review delivered
TIP
Why this kind of engagement pays off
Solo and small-team founders rarely have the signal to tell whether a contractor is underperforming until the bill is measured in months of lost runway. An embedded AI-native CTO catches exactly those problems — and the cost delta of one correction often covers the entire engagement many times over. Rising Seer has already demonstrated that math.