Cloud-based stock trading bot

I am looking to get a cloud-based stock trading bot done for myself, integrating two of my trading accounts and implementing on options strategy. Details are as below:

1. Strategy Overview
PCS is a dynamic intraday option selling automation designed to:
Harvest premium decay
Manage and hedge risk through recursive defense
Scale positions adaptively across multiple premium zones
Enable full intraday and overnight protection logic
Allow manual intervention when required.

The system must run fully automated based on user settings, while providing full manual override capability at any stage.

2. Full System Architecture
2.1 Core Functional Blocks
Parent creation
Child creation and recursive defense
Zone upgrades
New Parent creation based on dynamic triggers
Lot tracking, margin management
Dynamic zone management
Buy far OTMs for margin hedging
Optional scalping of bought options
IV and Weekday dynamic settings adjustment
EOD Protection and Overnight carry logic
2.2 Account Integration & 2FA Handling
Two Trading Accounts Connection:
E.g., Zerodha + another broker.
System must be able to connect to both accounts simultaneously.
Must support synchronous trade placement across both accounts.
Trade allocation ratio configurable if desired (e.g., 50:50 across accounts).
2FA Login Management:
System must handle broker 2FA securely:
TOTP (Time-based OTP),
API key/token based login sessions.
Login renewals must be automated at strategy start every morning.
Alert if login fails.
2.3 UI/UX Modules and Pages
ModuleFeaturesDashboard (Live Control Room)Live positions (Parents, Children) chain-wise viewLive margin usage graphNet MTM (profit/loss)Lot tracking (per zone, per chain)IV level, Day settings, Current Family Mode displayEmergency Global Pause/Kill buttonsMarket feed (basic Nifty data, IV data)Parameters & Configurations PageInitial lot size inputSubsequent lot size inputLoss % thresholds for defenseProfit booking thresholds for ChildrenZone definitions (absolute/percentage)Upgrade settings (max upgradeable Parents)Scalping settings for bought OTMsIV bucket sensitivity settingsDaywise defaults override optionFar OTM buy settingsEOD protection settingsManual Intervention Control PageModify open lots manually (adjust thresholds, switch family mode)Force square-off individual Parents/ChildrenAdd manual new Parent pairAdjust scalping lots manuallyLock specific lot from creating Children (freeze Parent)Reports PageLive log of trades (Parents created, Children created, exits booked)Zone movement logMargin usage logScalping log (buying scalping trades with timestamps)EOD full summary report (pnl, margin usage, chain summaries)

3. Core Trading System Functionalities
3.1 Full Automation of Trade Management
Auto place all entries (Parents, Children, Hedges) as per rules,
Auto manage defense, upgrades, exits,
Auto margin checks before every new trade,
Auto updating internal state based on live price feed.
3.2 Manual Overrides (Must Always Be Available)
Global Pause: Pause entire PCS system (no new trades until resume).
Global Kill: Force square-off all trades and stop strategy.
Manual Force Exit: Exit any single lot manually from dashboard.
Manual Additions:
Add manual Parent,
Add manual hedge,
Manual override zone for Parent,
Manual change of Family Mode (ATM to Pyramid mid-day if required).
3.3 Margin & Risk Management
Auto calculation of live margin usage per account.
Alert system when margin reaches user-set thresholds (e.g., 80%, 90%).
Pre-check before placing any new lot whether margin is available.
Smart scaling down:
Auto reduce lot size for new Children if margin is stretched.
3.4 Far OTM Buy Scalping Logic
Track bought options for scalping targets.
Auto sell them if target profit or stop-loss hit.
Track which bought options are still needed for margin safety.
Never allow scalping that disturbs minimum margin protection requirements.
4. Security and Safety Layers
AspectHandlingLogin SecurityAPI Key, TOTP for 2FA, session renewal checksTrade FailuresRetry logic + alert on repeated failuresMargin SafetyPre-trade checks mandatory; no margin breach allowedEvent Throttling1–2 second batching under high-speed movesKill SwitchAlways available as manual overrideLogsFull trade, event, and error logs maintained locally + cloud 

5. Deployment 
Deploy as web-based app (private cloud, private IP access),
Mobile UI optional (can be built after desktop),
Modular microservices architecture:
Trading engine,
Monitoring engine,
Scalping engine,
UI engine independent.
Use modern tech stack:
Backend: FastAPI / Node.js / Django,
Frontend: React / Next.js / TailwindCSS,
Database: PostgreSQL / MongoDB for fast object tracking.
6. Developer Notes
✅ Must modularize each PCS logical unit clearly:
Lot Handler (Parent/Child creation and monitoring)
Zone Handler (zone upgrade/exit logic)
Margin Handler (margin monitoring, auto scaling)
Account Manager (2FA, connection, trade execution split)
UI Handler (live visual chain maps, manual overrides)
Scalping Handler (for far OTM buys)
Risk Guard (kill switch, margin alerts, retries)
Separate logging for:
Trade activity,
Scalping activity,
Errors/failures,
Manual intervention logs.

Extra Setup to Make It Super-Responsive
Use WebSocket connections for LTPs, not polling REST APIs.
Keep order placement threads separate from LTP listener.
Use async, multi-threaded architecture (Python’s asyncio or Node’s event loops).
Pre-load strike lists and cache data — avoid calculating from scratch every millisecond.
Implement priority queues for orders — urgent ones get fired first.
AWS Mumbai server + clean code architecture + WebSocket data + margin hedges = perfect combo 
Do need consistent sub-1.5 second reactions, live monitoring, and clean execution.
WebSocket-based LTP fetching (not slow REST API polling).
Separate processing threads for market data, order decision, and order placement.
Broker APIs that can reliably complete order in 300–700 ms range (Zerodha, Shoonya usually are okay).
No big bottlenecks inside my code (no heavy loops, no database lag).
 ... Show more