Fintech software development
Software where a slow path or a missed check costs money.
Acasa Labs builds fintech software where speed and correctness are the product: trading systems, risk checks, realtime dashboards, and the billing behind subscription products. Our main fintech build is a private options trading platform for a confidential client on an Indian exchange, with tick capture, backtesting, paper and live execution, and a control dashboard.
The problem
In finance, the edge cases are the product.
Most software can be a little slow or a little wrong and nobody notices. Fintech cannot. A market-data burst that blocks the order path, a risk limit that does not hold, a payment recorded in one system and not the other: each one is money. Teams building these systems need engineers who measure latency instead of guessing it, put risk checks in front of every order, and treat reconciliation as a feature rather than a chore.
What we build
- 01
Trading and execution systems
Market-data capture, strategy execution, and order routing, with the same code path in research, paper, and live.
- 02
Risk controls
Position caps, per-bar guards, kill switches, and independent gates in front of live orders.
- 03
Realtime dashboards
Market state, positions, and system health streamed to a web dashboard without touching the trading path.
- 04
Billing that reconciles
Subscriptions across the web and the app stores, multi-currency payments, and payouts that agree with each other.
From production
From the options trading platform.
Measured in microseconds
Options trading platformMarket-data decode runs at 2.5 µs p50 and 8 µs p99, and a risk check at 21 µs p50. Measured in tape replay, with latency instrumented across 12 stages.
Trading never waits for the dashboard
Options trading platformMarket state fans out through Redis and a WebSocket gateway, throttled per instrument. The publisher is bounded and drops when full, so if the gateway dies, trading does not notice.
A kill switch that keeps the data
Options trading platformThe kill switch stops new risk without dropping the market-data stream, and live orders sit behind several independent gates.
Research on the recorded data
Options trading platformTicks are archived in Parquet and queried with DuckDB and Polars, so research and backtests run on the data the system itself captured.
Honest trade-offs
What we have built, and what we have not
Our fintech production work is trading: market data, research, risk, and execution, for a client whose name, broker, and strategies we keep private. We have also built subscription billing, multi-currency payments, and creator payouts inside our own products. We have not built a lending platform, a payment gateway, or a core banking system, and we will not pretend otherwise. If that is your project, we will say what we can bring and what to look for elsewhere.
What we reach for
- Python
- Rust-core event engine
- FastAPI
- Redis
- WebSocket, SSE
- Postgres
- Parquet, DuckDB, Polars
- Prometheus, Grafana
Proof
Shipped, not pitched.
01Client work
Options trading platform
Microsecond market-data handling for index options, with risk checks that never sleep.
A private trading platform for index options on an Indian exchange: tick capture, research and backtesting, paper and live execution, and a control dashboard. Built for a client's internal use; details are anonymized.
Trading engine · Realtime gateway · Web dashboard
Questions
Fintech questions we get asked.
Can you name the trading client?
No. The client, the broker, and the strategies are confidential, and we keep them that way. We can walk you through the architecture and how it was measured.
Do you handle regulatory compliance?
We build the controls and audit trails your compliance team specifies, but we are engineers, not a compliance firm. Your legal and compliance advisers decide what the rules require.
Which brokers and exchanges can you connect to?
The platform we built uses a custom adapter for its broker's REST and WebSocket APIs. A broker or exchange with a documented API can be connected the same way, after we check its limits and order types.
How do you measure latency?
Stage by stage, with exact percentile reservoirs exported to Prometheus and Grafana with alert rules. On the trading platform, 12 stages are instrumented, measured in tape replay.
Do you build payment and subscription systems?
Inside our own products, yes: Stripe and RevenueCat reconciled in NoteTube, six-currency Razorpay billing in Kraftr, and Stripe Connect payouts for educators. We have not built a payment gateway itself.
Will you sign an NDA?
Yes, before you share anything sensitive. The trading platform appears on our site anonymized, with client, broker, and strategy details withheld.
Fintech software development
Building something where milliseconds or money are at stake?
Tell us what the system has to do and where it breaks today. We will come back with an architecture and a plan.