Available for contracts & consulting

I build systems that
replace manual work
and operate at scale.

Senior developer specializing in government workflow automation, algorithmic trading infrastructure, and applied AI. I ship production systems that reduce cost, eliminate bottlenecks, and run without supervision.

View Projects Get in Touch
10+ years production systems · Government · Trading · AI · Node.js · Python · Go
01 — About

I build systems that
replace manual work.

I'm a solo developer with deep experience building production systems across three domains: government workflow automation, algorithmic trading infrastructure, and applied AI. My work runs in live tehsil offices, on 24/5 trading servers, and in the hands of hundreds of operators and farmers across Maharashtra.

I don't build for demos or portfolios. Every system I ship has a measurable outcome — fewer staff needed, faster processing, lower cost per transaction. I own the full stack: requirements, architecture, implementation, deployment. No handoffs. No miscommunication.

When I take on a problem, the goal is always the same: make the system work reliably at scale, then make it redundant to human effort. That mindset has driven every project I've delivered — from biometric RTO hardware integrations and AI-powered contract analysis to custom UDP transport layers built from scratch in Go.

10+ Years in production
10 Production systems shipped
90% Avg. labor reduction across automation projects
02 — Stack

Tools I trust
in production.

Trading Systems

  • MetaTrader 5 Platform, EAs, execution layer
  • MQL5 Custom indicators & trade logic
  • Algo Trading Session-based, risk-managed strategies
  • Trade Copier Multi-account fan-out via Node.js

Backend & Infrastructure

  • Node.js Core runtime for most systems
  • Fastify High-performance REST APIs
  • Go Concurrent, low-latency services & networking
  • UDP / Sockets Custom transport & protocol design
  • PostgreSQL / SQLite Storage layer
  • Redis Caching & pub/sub

Automation & Bots

  • Puppeteer Headless browser automation
  • Playwright Cross-browser workflows
  • Chrome Extensions Manifest V3, content scripts
  • WhatsApp / SMS APIs Citizen notification systems

AI / Applied Logic

  • Python AI logic, signal processing, scripting
  • Rule-based NLP Pattern matching, clause detection
  • Workflow Automation Cost-optimized process design
  • Hardware SDKs Biometric & webcam integration
03 — Projects

Systems built.
Problems solved.

01

ContextGuard — AI Contract Analysis Tool

Python Rule-based NLP Applied AI

Problem

Most people sign contracts without reading them carefully. Hidden clauses — around liability, auto-renewal, data ownership, and termination penalties — go unnoticed until they cause damage. Legal review is expensive and inaccessible for most users.

Solution

Built a lightweight contract analysis tool in Python. Scans uploaded contract text using a rule-based NLP engine that flags suspicious or high-risk clauses via keyword patterns, structural heuristics, and contextual scoring. No heavy ML infrastructure — the logic is deliberate, fast, and fully auditable.

Impact

Identifies risky clauses in seconds. Output is plain-language risk summaries — not raw annotations. Applied AI where it actually works: precise pattern detection over structured text, without the overhead of a large model.

02

Tehsil Automation System — Government Workflow

Puppeteer Node.js SQLite Chrome Extension

Problem

Desk 1 and Desk 3 operators at tehsil offices were manually processing citizen forms across multiple government portals — verification, data entry, submission — each case taking 8–12 minutes. Hundreds of daily cases created chronic backlogs and high error rates.

Solution

Deployed a Puppeteer-based automation layer that handles form filling, multi-portal verification steps, and submissions from a local citizen database. A companion Chrome Extension injects workflow logic into live government portals, eliminating navigation overhead without modifying the portal itself.

Impact

Per-case processing time dropped from ~10 minutes to under 60 seconds. Staff dependency at affected desks reduced significantly. Error rate down 90%. Operators now handle 5x more cases per shift. Running daily in active tehsil offices.

03

Farmer Notification System — KYC & Relief Alerts

Node.js WhatsApp API SMS Gateway cron

Problem

Farmers in rural areas were making repeated trips to tehsil offices to check KYC verification status and disaster relief payment updates — often only to be told nothing had changed. No proactive notification system existed. The time and cost burden fell entirely on the farmer.

Solution

Built an automated notification pipeline that monitors case status changes and triggers WhatsApp and SMS alerts to registered farmers. Covers pending KYC actions, document requirements, and payment disbursement confirmations. Fully automated — zero staff involvement per notification sent.

Impact

Physical office visits for status queries reduced by ~90%. Improved transparency for farmers who previously had no visibility into their cases. Eliminated a whole class of service requests that previously consumed desk staff time daily.

04

Farmer Card System — Government Services Membership

Node.js Fastify PostgreSQL Payment Integration

Problem

Farmers repeatedly pay per-transaction fees for standard government documents — 7/12 extracts, form submissions, record lookups — throughout the year. No bundled access model existed. Every interaction was a separate cost and friction point.

Solution

Built a yearly membership system where farmers pay a flat annual fee to receive a Farmer Card. Benefits include free 7/12 extracts and zero-cost access to all standard government forms for 12 months. Backend manages membership state, entitlement validation, and renewal lifecycle.

Impact

Reduces per-transaction friction significantly. Encourages farmers to engage with government services digitally rather than through paid intermediaries. Creates a sustainable service model for operators while lowering the annual cost for end users.

05

MT5 Trade Copier — Multi-Account Infrastructure

MT5 MQL5 Node.js REST API

Problem

Managing multiple MT5 accounts from a single master strategy requires simultaneous execution across all accounts with minimal latency. Third-party copy services introduce lag, cost, and no control over execution behavior or lot sizing logic.

Solution

An MQL5 Expert Advisor on the master account pushes trade signals to a Node.js relay server via HTTP. The server fans signals out in parallel to all connected slave accounts in real time. Supports dynamic slave registration, configurable lot scaling per account, and automatic reconnection on socket drop.

Impact

Average copy latency under 80ms. Supports 10+ slave accounts per master. Six months of live operation with zero missed trades. Eliminated dependency on third-party copy services entirely — full control, zero recurring infrastructure cost.

06

EUR/USD Algorithmic Trading System

Python MQL5 MT5 Pandas

Problem

Discretionary trading is inconsistent — subject to emotion, fatigue, and missed setups outside active hours. A systematic edge needed to be codified and executed without human involvement across all active market sessions.

Solution

A Python strategy engine analyses price action, session context (London/NY overlap), and momentum signals to generate high-probability entry criteria. Signals feed into an MQL5 EA that executes on MT5 with ATR-based dynamic SL/TP, equity-tied position sizing, and session-aware trade filtering to avoid low-liquidity windows.

Impact

Backtested over 3 years of tick data with consistent positive expectancy across both primary sessions. Runs 24/5 on a dedicated VPS. Zero manual intervention. Built for long-term, reliable execution — not an experiment.

07

UDP Transport System — Custom Low-Level Networking in Go

Go UDP Android Socket Programming

Problem

Most communication systems rely on TCP or third-party tunneling services, which introduce overhead, latency, and external dependencies. Building a custom transport layer from scratch — with full control over packet flow, message framing, and routing behavior — requires working at the socket level without abstractions.

Solution

Built a UDP-based transport system in Go that handles direct system-to-system communication with custom message framing and packet flow logic. An Android client acts as a connected endpoint, communicating with the Go server over raw UDP sockets. The architecture explores proxy-like and tunneling patterns — routing traffic through a controlled relay without depending on any third-party service or protocol library.

Impact

Demonstrates practical depth in low-level networking: socket lifecycle, UDP datagram handling, and custom protocol design. Serves as a working foundation for future distributed systems that require direct, low-overhead communication channels. The Go server and Android client together validate the full transport path end-to-end.

04 — Experience

How I work.

⬡

Solo, end-to-end ownership

I own every phase: requirements, architecture, implementation, testing, deployment. No handoffs means no context loss, no delays, no ambiguity. One person who knows the entire system — from database schema to production config.

⬡

Production-first, always

I don't build prototypes and call them done. Every system I deliver is designed for real load, real edge cases, and real failures. Error handling, retry logic, logging, and graceful degradation are built in from day one — not added later.

⬡

Speed without shortcuts

I move fast because I've solved similar problems before. Proven patterns, no over-engineering, incremental delivery. Speed comes from experience and clarity — not from skipping reliability.

⬡

Cost and labor reduction as core metrics

Every system I build is measured by one question: does it reduce cost or increase throughput? Whether cutting man-hours at a government desk, eliminating trading infrastructure fees, or reducing farmer visits by 90% — ROI is always in scope.

⬡

Deep domain experience across multiple fields

Government workflow automation, algorithmic trading infrastructure, applied AI, and low-level networking — not surface-level familiarity, but years of production systems in each. I understand the real constraints, edge cases, and what actually works at the operator level.

05 — Contact

Have a system
that needs building?

Available for contract work, consulting, and long-term partnerships. If you have a real operational problem — government workflows, trading infrastructure, automation, or applied AI — I can scope it, build it, and own it end-to-end.

ceo@abhiform.com github.com/alexpatil