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 · 1 build slot at a time — currently open·What “fractional” means →
Verifiable, not claimed
Data
- 3
- systems running in production
- 4,912
- pipeline partitions at 99.9%
AI
- 28
- approval-gated workflows
- 42
- permission-gated tools
Trusted by teams at
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 real-time macro analysis to enterprise compliance to high-throughput financial data processing — each represents a different kind of data challenge.
Financial Analytics
Stockopedia
Designed and built the core data-engineering platform powering stock analysis for investors across Europe — high-throughput pipelines processing financial data across global markets.
Read the case study →Compliance / Internal Audit
Sehati
Built an end-to-end internal audit tool covering the full audit lifecycle — planning, risk assessment, workflow automation and real-time reporting for compliance teams.
Read the case study →Industrial / 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 →Products
Products I build and run
The consulting expertise is proven in production — through my own products, from Shariah-compliant investing to open-source infrastructure.
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.
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.
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.
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.
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. In one week you get 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 it doesn't surface fixes or savings worth more than the fee, you don't pay.
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. One senior engineer, delivered in weeks, priced against what the hire or the DIY quarter would really cost you.
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 — so an LLM can do real work without doing something you cannot undo.
About
A bit about me
I'm Arshad — a data engineer with over 15 years of experience building systems that process, analyze, and act on data at scale. Most of my career has been in financial services, including building the core data platform at Stockopedia, where I designed the engineering backbone that powers stock analysis for investors across Europe.
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?
I build two things for you: lean data platforms — pipelines, a warehouse, dashboards — and AI automation that runs in real operations with human approval gates. It is a solo practice run by Arshad Ansari, so you work directly with the person building the system, not an account manager.
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, 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?
Most projects start with the Data Platform Audit — $3,000 fixed, one week, guaranteed: if it doesn't surface fixes or savings worth more than the fee, you don't pay. If we go further, builds are scoped in weeks and priced against your real alternatives: a cloud-warehouse bill, or the cost of a senior hire — and the audit fee comes off the build price. Capacity is real, not marketing: two audits a month, one build at a time. The first step is always a free 30-minute scoping call.
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: 42 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.
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.
Read the 18-check teardownOr 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.
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]