Available for Senior Full-Stack, Backend, or Automation Platform roles

Full-Stack & Backend Systems Engineer

Building robust Spring Boot APIs, Next.js applications, and platform automation engines.

4+ years architecting high-throughput JVM backends, deterministic Playwright automation engines, and modern Next.js/React full-stack interfaces. Production experience spanning enterprise retail platforms (Reflexis / Zebra) to distributed ERP automation.

Primary Backend & APIs

Java 21Spring Boot 3REST APIsMicroservicesPostgreSQLMySQLDB2

Frontend & Full-Stack

Next.js 15/16ReactTypeScriptTailwind CSS

Platform Automation & Tooling

PlaywrightChrome MV3Headless Node DaemonsDocker

Location & Mobility:Prayagraj / Remote (Open to Bangalore, NCR)

Direct Inquiry:shivambhaipatel1997@gmail.com

Download Resume (PDF)↓ATS Plaintext View↗

01PRODUCTION SYSTEMS

Full-Stack Applications, Distributed Backends & Automation Engines

Selected Work

Production software built under real performance, concurrency, and reliability constraints.

Enterprise Backend · Deterministic Skia CompositionLive Microservice ↗

PracharFlow

Low-latency localized rendering engine serving automated bots. Replaced unpredictable image models with a deterministic parametric canvas engine built in Java 21 and Spring Boot 3 using Skia (Skija). Achieved <150 ms layout compilation for complex Devanagari and Gujarati scripts with zero GPU overhead.

Generative diffusion models hallucinate Devanagari and Gujarati ligatures, distort official brand and political party logos, and take 8+ seconds per generation. Campaigns required guaranteed typography accuracy and sub-second generation.

Rejected generative image pipelines in favor of a deterministic parametric canvas engine. Built in Java 21 / Spring Boot 3 using Google's Skia (via Skija), rendering structured layout trees with typed slot constraints directly to PNG buffers.

Consequence & Verified Outcome

Pixel-perfect regional typography and exact logo rendering in under 150ms with zero GPU infrastructure costs, serving automated Telegram bot workflows.

Java 21Spring Boot 3Skija / SkiaPostgreSQLRedis

Case study →

PracharFlow Rendering PipelineVector Core

Hover stages to trace pipeline execution

Engineering Trade-off AnalysisGenerative Diffusion vs. Skia Engine

Dimension	Generative Diffusion	Deterministic Skia
Devanagari / Gujarati	Hallucinates ligatures & broken conjuncts	Native HarfBuzz font shaping (100% accurate)
Brand & Emblems	Distorts official party logos & vectors	Bit-exact rasterization into typed layout slots
Execution Latency	8–15s per image on GPU infrastructure	<150ms on standard CPU JVM container
Reproducibility	Stochastic output across random seeds	100% deterministic, bit-exact rendering

Flagship Architecture · Multi-Agent Control PlaneInteractive Sandbox ↗

Workflow Studio

Orchestration engine for autonomous developer tasks and code execution across shared codebases without state collision or branch drift. Built on an isolated task-dispatch engine backed by a local SQLite DAG.

Architecture Boundary & Topology

Hover or tap any module to inspect contract isolation

Strict tsconfig Boundary

VS Code ExtensionEditor Surface

TreeViews & IPC dispatch via VS Code API

CLI ToolTerminal Engine

bin/workflow-studio.mjs · Headless runner

↓ IPC Bridge

Decoupled Engine

@workflow-studio/core

Task DAG · SQLite Engine · IPC Dispatcher

types: ["node"] (headless Node runtime)

↓ Subscriptions

Dashboard SPATelemetry UI

Vite / React state client over local HTTP

TradeSense DeskDownstream App

link:../workflow-studio/packages/core strict import

Autonomous Multi-Agent SDLC Cycle

1. Task DAGIngestion

Schema-Validated Task Decomposition

2. Worker AgentExecution

Isolated Git Workspace

3. Reviewer GateReviewer Gate

AST & Downstream Contract Assertions

4. Sync & PRDelivery

State Re-verification & Sync

// ACTIVE CONTRACT:@workflow-studio/core

Strict Node.js types: ["node"] · Zero UI dependencies · SQLite persistence

The Architectural Challenge

The engine originally lived coupled inside a VS Code extension, making it impossible to drive from CLI scripts, CI test suites, or headless worker sessions without booting a full editor instance.

The Engineering Solution

Decoupled runtime logic into an isolated headless Node.js core engine (@workflow-studio/core) backed by an explicit SQLite task DAG using strict compile-time tsconfig boundaries. Implemented Kahn's algorithm for topological task ordering with strict cycle detection, and distributed task leases with monotonic fencing tokens to reject stale writes.

Node.jsTypeScriptReactNext.jsSQLiteTask DAG

Read full case study →

Platform Automation · Two-Tier Isolation & ReplayLive Engine ↗

FlowTrace

Deterministic regression recording and replay engine for dynamic enterprise web apps like Oracle Fusion ERP. Solves brittle selector failures across quarterly vendor updates using a two-tier verification engine (Chrome MV3 + headless Node.js Playwright daemon) with 4-step self-healing fallbacks.

The Constraint & Trade-off

Enterprise ERP systems regenerate DOM element IDs across sessions and quarterly patches (`#btn-submit-8942` → `#btn-submit-3180`). Standard test runners break immediately or report false greens without validating that the business operation succeeded.

Architecture Decision

Built a two-tier verification engine combining Chrome MV3 telemetry with a headless Node.js Playwright daemon. Designed a 4-step self-healing fallback (ARIA accessibility tree → contextual parent attributes → action text) with state mutation assertions (verifying modal dismissals and network responses before confirming step execution).

PlaywrightTypeScriptNode.jsChrome MV3Enterprise ERP

Case study →

FlowTrace Architecture PipelineTwo-Tier Engine

Hover steps to inspect boundary contracts

Selector Drift Recovery Pipeline

Handling vendor quarterly DOM ID mutations in enterprise ERPs

Runtime Oracle Fusion DOM (Vendor Patch):

<button id="btn-submit-3180"
  data-ora-ctx="po_approval_flow"
  aria-label="Approve Purchase Order">
    Approve Order (₹ 4,80,000)
</button>

* Quarterly patch mutated ID from #btn-submit-8942 → #btn-submit-3180.

Tiered Resolution Sequence:

✓ Tiered Fallback Matches Element:

Check ARIA accessibility tree: aria-label="Approve Purchase Order"
Match parent context: data-ora-ctx="po_approval_flow"
Normalize action text: "Approve Order"
Verify state change (modal dismiss + toast) → writes patch to JSON contract.

02APPLICATIONS & GATEWAYS

Full-Stack Applications, Unified Gateways & Analysis Terminals

Also Running

Full-stack web applications, unified enterprise API gateways, and market intelligence terminals.

Project	Architecture / Summary	Stack	Status
TradeSense	Multi-agent market research and regime analysis terminal built on top of Workflow Studio.	TypeScriptNext.js@workflow-studio/core	Live Terminal ↗
Enterprise ERP Connector	Unified API gateway interface and reference architecture modeling unified query paths across heterogeneous enterprise systems (Oracle Fusion, SAP, Cognos, Power BI).	Java 21Spring Boot 3REST APIsMicroservices	Live Prototype ↗
DealDekho	Multi-marketplace price intelligence platform across Amazon, Flipkart, and Croma with Next.js full-stack architecture, resilient price normalization pipelines, and PostgreSQL persistence.	Next.jsReactNode.jsPostgreSQLDistributed Scraping	Live Platform ↗/GitHub ↗

03ENGINEERING CAREER

Enterprise Backends & Systems Architecture

Experience

Independent Software Consultant — Full-Stack & Systems Automation

Client Advisory & Independent Engineering

Apr 2024 – Present·Remote, India (Prayagraj)

Enterprise backend systems, high-concurrency Spring Boot microservices, ERP integration pipelines, and deterministic automation engines across commercial client engagements.

↳Retained & Contract Engagements:

FirstCron Services Pvt Ltd(Product: SyntraFlow)

Location: Delhi, India · Mar 2026 – Present

Current client · Ongoing

Engineered SyntraFlow, an enterprise Oracle Fusion ERP automation and regression verification platform operating across 50+ business entities at a single enterprise client.

Architected and deployed end-to-end regression automation infrastructure for Oracle Fusion ERP across 50+ business entities, building a three-tier execution engine (MV3 + Node runner + Playwright) cutting regression cycles by 40%.
Built a high-throughput Playwright execution engine with deterministic DOM state assertion, self-healing selectors, and automated SOC2 compliance audit trails.
Architected Spring Boot API gateway layers with fine-grained role-based access control (RBAC) and multi-tenant persistence.

Java 21Spring Boot 3TypeScriptPlaywrightReactMySQLAWS

SamMegh Technologies

Location: India · Apr 2024 – Feb 2026

Contract engagement · Completed

Enterprise systems consulting, database query optimization, and Spring Boot backend performance tuning.

Diagnosed database execution bottlenecks and refactored Spring Boot query paths, reducing end-to-end API response latency from ~800 ms to under 300 ms. Implemented query plan indexing and connection pool tuning for high-throughput services.
Implemented database query plan indexing, statement caching, and refactored entity fetching to eliminate N+1 queries across core microservices.
Hardened REST contracts and persistence boundaries with strict DTO validation and idempotency keys.

JavaSpring Boot 3MicroservicesPostgreSQLHikariCPREST APIs

Software Engineer — Platform & Enterprise Retail

Reflexis Systems (Zebra Technologies)

Nov 2021 – Mar 2024·Pune, India

Core engineering on enterprise retail workforce management platforms servicing 150K+ daily active store associates across 400+ enterprise retail store locations under strict sub-100ms API SLAs.

Engineered and scaled enterprise retail workforce management platforms serving 150K+ daily active store associates across 400+ locations under strict sub-100ms API SLAs.
Refactored monolithic legacy services into domain-bounded Spring Boot microservices with isolated DB2 persistence schemas.
Maintained high-availability retail store services; diagnosed memory bottlenecks, analyzed heap dumps, and tuned G1GC parameters to increase transaction throughput by 30% during retail peak surges.
Overhauled automated Jenkins CI/CD deployment pipelines, cutting production release cycles from bi-weekly to under 2 days.
Expanded integration and unit test coverage across core scheduling services, reducing post-release defect tickets by ~25%.

JavaSpring BootIBM DB2MicroservicesJVM TuningEclipse MATJenkinsDocker

↳Academic Background & Credentials

Post Graduate Diploma in Advanced Computing (PG-DAC)

Centre for Development of Advanced Computing (C-DAC), ACTS Pune

Mar 2021 – Sep 2021·Pune, India

Bachelor of Technology (B.Tech) in Electrical Engineering

Institute of Engineering & Technology, M.J.P. Rohilkhand University

Aug 2016 – Nov 2020·Bareilly, India

04ARCHITECTURAL DECISIONS

Trade-Offs & Rules of Thumb

Engineering Trade-Offs

Practical architectural heuristics and trade-offs drawn from building high-throughput JVM backends, browser automation suites, and developer platform tooling.

Compile-Time Boundaries over Runtime Discipline

Enforce architectural separation where the compiler can catch violations rather than trusting team conventions or documentation.

Evidence:In Workflow Studio, packages/core/tsconfig.json scopes types strictly to Node. Any accidental import of VS Code UI or editor APIs into the core engine triggers an immediate build-time error rather than a runtime failure.

State Verification over False Greens

A fast, loud failure is cheap to fix; an unverified false green that silently reports success masks catastrophic defects.

Evidence:In FlowTrace, the replay engine refuses to mark an automation step complete until the destination DOM state satisfies strict assertion rules. Replaying a save button click requires verifying subsequent modal state mutation.

Deterministic Canvas over Generative Hallucination

When business requirements demand 100% typography accuracy and brand fidelity, choose deterministic code composition over stochastic AI models.

Evidence:In PracharFlow, regional campaigns require exact political symbols and Devanagari/Gujarati script rendering. We built a Java 21 / Skia compositor with typed layout slots, delivering sub-150ms rendering with zero GPU overhead.

Upstream Aggregation over Adversarial Scraping

Avoid brittle, high-maintenance anti-bot arms races by pairing upstream price comparison APIs with verified client-side telemetry.

Evidence:In DealDekho, catalog indexing pairs upstream aggregator APIs with price verification telemetry from the ShopLens browser extension rather than maintaining fragile direct scrapers against marketplace bot guards.

05DIRECT COMMUNICATION

Availability & Fast Handoff

Contact

Available for Full-Time Roles & Technical Consulting

Timezone: IST (UTC+5:30) · Flexible US/EU Overlap

Looking for systems, backend engineering, or developer platform roles, as well as technical consulting on enterprise automation engines, multi-agent SDLC systems, and high-throughput JVM backends.

Direct Technical Conversation

Book a 20-min Architecture / Intro Call

Discuss system requirements, task decomposition, JVM bottlenecks, or browser-automation regression suites directly.

Initiate Technical Discussion↗

Asynchronous Communication

shivambhaipatel1997@gmail.com

Detailed technical queries, PRD review requests, or formal RFPs. Direct response typically within 24 hours.

Send Direct Email ↗

LinkedIn↗/GitHub↗/Resume↗