Work with Ravi
I run three production AI SaaS solo (Prism, Citare, BatchWise) and write about the actual mechanics of LLM FinOps. If your AI spend is meaningful and you want practitioner-level help — not a Big-4-style slide deck — here's how to engage.
Who I work with
Indian mid-market companies (typically ₹50 crore - ₹2,000 crore revenue) with material AI / cloud spend, and solo founders or small engineering teams running production AI products with monthly LLM bills above $10K. If your AI bill is under $1K/mo, you don't need me yet — read the LLM FinOps explainer and the $5/month production stack first.
What I do
1. AI Systems Review
Six-domain independent review across AI spend attribution, vendor consolidation, model selection, cloud architecture for AI workloads, Indian regulatory overlay (SEBI / ICAI / DPDP / RBI), and internal controls. Output: written report with findings, quantified optimisation roadmap, governance framework, and executive deck.
Not a statutory audit. Advisory in nature; you retain decision authority.
2. AI Spend & Tax Optimisation
Tax-recovery diagnostic across five sub-domains: GST input credit on AI / cloud imports, Section 195 TDS reconciliation with DTAA relief, Reverse Charge Mechanism on imported services, capex vs opex classification under Ind AS, transfer pricing on cross-entity AI cost allocation. Output: quantified recovery roadmap your existing CA can execute against.
Not a tax opinion or filing service. Diagnostic only; implementation runs through your existing CA firm.
3. LLM FinOps Coaching
Embedded coaching for an engineering or finance team implementing LLM FinOps discipline — cost attribution architecture, caching, model routing, retry hygiene, governance — without hiring a full-time specialist. Two calls per month plus async review of code and dashboards.
Fit: teams with $10K-$100K/mo LLM spend who want capability inside the team, not a vendor relationship.
Why bespoke pricing
Engagement scope varies too much across entities for fixed tiers to be honest. A multi-BU group with ₹50 crore annual AI spend and EU subsidiary exposure has fundamentally different requirements than a single-entity SaaS with ₹3 crore spend. The free 30-minute discovery call produces a written scope proposal with the exact price before any commitment. No pressure. If we're not the right fit, I'll say so.
How to start
Two paths, both free:
- Book a 30-min call — for when you'd rather talk through the situation than write it down.
- Send context first via the intake form — for when you want me to read up before we talk, or when async written response is a better fit than a call. I read every form personally and respond within 24h Mon-Fri.
About me
I'm Ravi Patel. I built and ship three production AI SaaS solo — Prism (an OpenAI-compatible AI gateway competing with Portkey and Helicone), Citare (AI search visibility tracker across AIO / ChatGPT / Perplexity / Gemini), and BatchWise (Indian compliance marketplace). I write practitioner notes on the work at rikuq.com.
The advisory practice exists because of a structural gap in the market: technologists deep enough to optimise LLM infrastructure usually don't do finance work, and finance professionals deep enough to handle Section 195 / GST RCM / Ind AS classification usually don't understand AI infrastructure well enough to optimise it. I sit in that gap and have shipped enough production AI to know what actually breaks vs. what only looks scary on a slide.