Hikmah Technologies
Production AI on your own data.
I'm Arshad Ansari. I build the data platforms and the approval-gated AI workflows that run on them — self-hosted, internals public — instead of the $150–200k senior hire you'd spend six months landing, or an AI agency that can demo a workflow but has never run one against production data.
I run all of it in production myself, on my own data, before any of it touches yours. Three live systems — so you can check the work before you pay for it.
2 audit slots a month · one full-time engagement at a time — currently open·What “fractional” means →·How I ship AI workflows →
Verifiable, not claimed
Data
- 3
- systems running in production
- 4,912
- pipeline partitions at 99.9%
AI
- 40
- automation workflows
- 52
- permission-gated tools
Who this is for
You probably recognise at least two of these
- →You have more data than a spreadsheet holds, and nobody whose actual job it is.
- →The warehouse bill is growing faster than your usage, and no one can say why.
- →Two dashboards answer the same question differently, and both get quoted.
- →Data work keeps landing on a product engineer who did not sign up for it.
- →A pipeline broke last month and you found out from a customer.
- →There are three AI pilots in the building and none of them has survived contact with production.
- →Someone has already wired an LLM to something that matters, and everyone is quietly nervous about it.
- →You need this working in weeks — not after two quarters of hiring.
If none of those land, you probably don't need me yet — and I'd rather say so here than on a call.
Your real alternatives
What you'd do instead
Hire a senior data engineer
$150–200k/yr + 3–6 monthsYou are hiring someone to invent a platform — and if you have no data lead, you cannot evaluate the candidates.
Give it to your product engineers
A quarter of roadmapThey learn data engineering on your production data, and the platform ends up shaped like whoever was free that sprint.
An agency or offshore team
$15–30k/monthYou get throughput, not judgement — and every rotation takes the context with it.
Buy the managed stack
A meter that compoundsFastest way to have a warehouse at all, and the reason most of the bills I get called about look the way they do.
An AI agency or automation shop
A demo, then a handoverThe demos are genuinely good. Ask how many of their workflows have run unattended against production data for a year, and who owns it the week after handover.
Wire n8n or Zapier to ChatGPT yourself
A weekend, then upkeepThe right way to find out in a week whether the idea works at all — and fine until the first silent failure reaches a customer, because nobody owns the retries, the guardrails or the audit trail.
Me, fractionally
$3,000 to startOne senior engineer who builds and runs the whole stack — the pipelines and the AI workflows on top of them — delivered in weeks and handed over documented, so you can hire into a working platform later instead of hiring someone to invent one.
If you genuinely have 40 hours a week of data work, hire the full-timer. I'll tell you that on the call.
Selected Work
Systems I've designed and built
From a UK capital-markets research platform to enterprise compliance — systems where the numbers have to be right.
Capital Markets
UK capital-markets research platform
Architected the core data system behind a UK research platform serving investors across Europe — high-throughput pipelines reconciling financial data across global markets.
Read the case study →Compliance / Internal Audit
Sehati
Built an internal-audit management platform for enterprise compliance teams — planning, risk assessment, fieldwork and findings, with every step recorded and reportable in real time.
Read the case study →The Lab
Live systems that prove the claims
Not a product catalogue — each of these runs its own business on its own domain. Here, each one is evidence for a claim this site makes.
Live at aeocue.com, ansaar.in, domainposture.com, quantamentry.com, koyracloud.com.
aeocue
Get named when your customers ask an AI
A done-for-you AI-answer visibility service for local businesses in Navi Mumbai. Every week it checks whether ChatGPT, Perplexity and Google's AI answers name the business, fixes what stops them, and sends the before/after receipts.
What it proves: the evidence discipline, pointed at AI answers — every claim is a dated before/after receipt, not a dashboard number.
Ansaar
AI-powered Shariah-compliant investing platform
A trading and investment platform for Indian retail investors with a Shariah-compliance focus — equity, crypto and ETF screeners, institutional "whale-watcher" signals, market-regime classification and ML-driven predictions.
What it proves: the data-platform claims — the 4,912 pipeline partitions at 99.9% on this site are this product's backbone.
Domain Posture
Audit-grade domain hygiene reports
Signed, hash-chained PDF evidence packs covering DNS, email authentication (SPF / DMARC / DKIM), TLS, security headers and CORS — defensible reports you can hand to auditors and customers.
What it proves: audit-grade evidence discipline — signed, hash-chained reports, because "trust me" is not a compliance answer.
Quantamental
Global macro & central-bank credibility analytics
A macro-analysis platform scoring central-bank credibility and tracking macro fundamentals across 171 countries — built on ten free public data sources (World Bank, IMF, ILO, BIS, FRED, OECD and more) and refreshed daily.
What it proves: LLM analysis over a governed warehouse, in production, refreshed daily.
KoyraCloud
Open SourceOpen-source self-hosted PaaS for Docker Swarm
An open-source (AGPL-3.0) platform-as-a-service that brings a Vercel/Render-style push-to-deploy experience to your own infrastructure — image builds, auto-TLS subdomains, persistent storage, workers and cron jobs on Docker Swarm.
What it proves: right-sized self-hosted infrastructure — "near-zero marginal cost" is a running PaaS, not a slogan.
Featured · Open Source
MITAEGIS
A self-hosted, flow-first personal AI orchestration platform
A small fleet of named agents run durable workflows over my own data — tasks, money, knowledge, homelab alerts — and ask for a decision only when they actually need one. It is the system I run my own days on, now open source for anyone to fork.
- →Durable, flow-first automation on Temporal
- →One human-in-the-loop primitive for every decision
- →Local-LLM-first, self-hosted, bring-your-own everything
Services
How I can help
Data Platform Audit
A fixed-price teardown of your data stack: 12–20 hours of senior work over one week — a ranked fix-list, a cost model comparing your spend to a lean rebuild, and a small proof-of-concept on your own stack. $3,000, and if day one shows no credible opportunity, you pay nothing.
Data Platform Build
I build the whole path — from your source systems into a warehouse (DuckDB, ClickHouse or Postgres) and out to dashboards your team can trust. $15–40k depending on scope, delivered in weeks — against the $150–200k senior hire you'd otherwise make.
AI Automation Build
I build AI automation that runs in production, not in a demo: approval-gated agents, document-to-table pipelines, and natural-language-to-SQL with guardrails. $10–25k per bounded process, plus a monthly retainer to run and improve it.
About
A bit about me
I'm Arshad — fifteen years building data platforms where the numbers have to be right; most recently the architect of a UK research platform's core data system.
I'm the author of Local-First Analytics (now on Amazon), a Databricks technology partner, and I run Hikmah Technologies as a focused consulting practice — helping companies get their data infrastructure right the first time.
FAQ
Common questions
What does Hikmah Technologies actually do?
One thing: production AI on your own data. I build the data platform — pipelines, a warehouse, dashboards — and the AI workflows that run on top of it, every consequential action gated by a human approval. It is a solo practice run by Arshad Ansari, so you work directly with the person building the system. The products on this site are the lab: live systems that prove the claims, not a separate business.
How do I know you can actually build this?
Because you can check before you hire me. I run three systems in production: AEGIS, an open-source AI automation platform you can read line by line at github.com/hikmahtech/aegis; Ansaar, a trading data platform live at ansaar.in with 4,912 pipeline partitions backfilled at 99.9%; and Quantamentry, a macro-data product at quantamentry.com scoring 171 countries daily. Most consultants show logos — I show you the source.
Do I need to hire a data team?
No — that is the point. I design, build and run the whole stack solo, from ingestion to warehouse to dashboards, and hand it over documented. A senior data engineer is a six-figure annual commitment and months of hiring; a scoped build with me is delivered in weeks.
Why not just hire a full-time data engineer?
If you truly have 40 hours a week of data work and $150–200k plus months to hire, you should. Most teams that call me have neither: the work is lumpy — heavy while the platform goes up, light once it runs. A fractional senior engineer gives you the judgement without the salary — $3–4k a month for one day a week, advisory only — and everything I build is documented and handed over, so you can hire into a working platform later instead of hiring someone to invent one.
How do engagements and pricing work?
Everything is priced in USD, in the open. The Data Platform Audit is $3,000 fixed: a free 30-minute scoping call first, I only recommend the audit after seeing your cost export, and if day one shows no credible opportunity, you pay nothing — 12–20 hours over one week, with the fee credited in full toward any follow-on build. Builds run $15–40k against the $150–200k senior hire; AI workflows are $10–25k per bounded process plus a monthly retainer; fractional advisory is $3–4k a month for one day a week, three-month minimum. Capacity is real, not marketing: two audits a month, one full-time engagement at a time.
Is it risky to let an AI agent take actions in my business?
It is, if the agent has free rein — so I do not build it that way. Every action that matters routes through a human approval gate before it happens, and everything is logged and traceable. AEGIS, my open-source platform, works exactly this way: 52 permission-gated tools, every human handoff a single record you approve or reject. You can read the code.
My cloud-warehouse bill keeps growing. Can you help?
Usually, yes. A lot of warehouse spend comes from paying per query for workloads that fit comfortably in DuckDB, ClickHouse or Postgres on right-sized infrastructure. The Data Platform Audit maps exactly where your money is going and what a leaner setup would cost — that is often where the audit pays for itself.
More buyer-mechanics questions — ownership, timezones, invoicing — are answered on the FAQ page.
Not ready to talk?
Take the audit and run it yourself
The 18 questions I work through in a $3,000 audit are published in full — cost, correctness, reliability, and whether you can safely change any of it. With what a bad answer sounds like for each. No signup, no gate; most teams find something expensive in the first hour. And for the AI half: the 17 questions I ask before an LLM workflow is trusted in production — who can stop it, what happens when it is wrong, what it really costs.
Or just stay in touch
I write up what breaks and what it costs — pipelines, warehouse bills, and the things that only show up in production. A few emails a month.
No sequence, no pitch deck. Reply 'stop' once and you're off — it reaches me, not a queue.
Indian Businesses
Based in Navi Mumbai
I also build for Indian businesses — lead-generation websites and business systems for manufacturers, traders and firms. WhatsApp is the fastest way to reach me.
WhatsApp me · +91 72084 61733Industrial / B2B
VC Solutions
Designed and shipped a fast, SEO-focused marketing and lead-generation website for an MPCB-licensed industrial solvents and scrap-metal trading business in Maharashtra and Gujarat.
Read the case study →Real Estate
Manas Realty
Built a modern marketing website for a real-estate firm operating since 1991 — project showcase, services and enquiry capture, with privacy-friendly analytics.
Read the case study →Need the data platform, and the AI that runs on it?
I take on a small number of projects at a time so I can do them well. The first step is a free 30-minute scoping call — most builds start with a $3,000 fixed-price audit whose fee comes off the build.
Book a free scoping callOr email [email protected]