Portrait of Ankit Verma

Backend, Automation, Internal Platforms

Building reliable systems that turn manual operations into repeatable software workflows.

I design APIs, workflow automation, enterprise integrations, and internal tools that help teams move faster with less manual coordination. My work sits at the intersection of backend engineering, business process automation, and operational clarity.

3+ years of experience. Software Development Engineer – Backend & Integrations at Gigmo Solutions Pvt Ltd, June 2023 – Present.

Download Resume LinkedIn
Core Work API services and integrations
Automation Workflows for HR, MIS, and operations
Platforms Power Apps, Dataverse, and data pipelines

About

From repetitive business tasks to dependable internal systems

I specialize in backend engineering and automation for teams that depend on accurate workflows, timely data, and traceable operations. That usually means taking a process that is spread across spreadsheets, email follow-ups, and manual approvals, then rebuilding it as a cleaner system with APIs, scheduled jobs, or Power Platform applications.

The work I enjoy most is the kind that improves day-to-day execution: payroll flows that are easier to observe, attendance pipelines that reconcile data automatically, approval systems that stop depending on inbox chaos, and business tools that give internal teams a more reliable way to operate.

How I Add Value

A blend of engineering depth and business workflow understanding

Automation Architect

I map operational bottlenecks and redesign them into structured flows with clear handoffs, fewer manual dependencies, and better visibility.

API Builder

I build backend services that handle integrations, validation, orchestration, and status tracking so internal processes can scale more cleanly.

Workflow Engineer

I use Power Apps, Power Automate, Dataverse, and supporting backend logic to turn approval chains and operational routines into maintainable systems.

Data Pipeline Builder

I create ETL style pipelines that clean, normalize, and reconcile data from different sources so teams can work from more dependable outputs.

Technology Stack

Tools I use to build backend and automation systems

Python
FastAPI
Flask
Pandas
Docker
Redis
Dataverse
Power Apps
Power Automate
SharePoint
Microsoft Graph
REST APIs

Selected Backend Work

Projects focused on reliability, integration, and operational automation

External Integration

Secure Callback Integration

Built a backend service for receiving encrypted transaction updates and document uploads from external systems. The goal was to turn third-party callbacks into a structured, validated workflow that internal teams could rely on without manual intervention.

  • Validated encrypted payloads before downstream processing
  • Handled document ingestion and workflow-safe status updates
  • Handles ~300-400 encrypted transactions per month

Data Reconciliation

Biometric Attendance Automation

Designed an automated reconciliation pipeline for biometric logs, leave data, and attendance records. The focus was accuracy, repeatability, and reducing the amount of manual cleanup needed each cycle.

  • Pandas-based ETL logic for normalization and rule processing
  • Spreadsheet ingestion and transformation workflows
  • Processes 100-120 employee records per day
  • Eliminates 2-3 hours per day of manual attendance cleanup

Service Modernization

API Driven Payroll Processing

Problem: Payroll-related workflows were tied to opaque cron-driven execution and were difficult to observe when support teams needed status or troubleshooting context.

Approach: Moved execution into API-oriented services with FastAPI endpoints for trigger and status driven processing, giving operations teams clearer orchestration and integration points.

Result: Processes 400-500 payroll records per run across 3-4 monthly cycles, with ETL batches handling 40k-50k records per batch.

Scheduled Automation

Payment Automation Engine

Problem: Recurring payout workflows depended on slow manual handoffs, stretching payout processing turnaround to about 5 days.

Approach: Extended payout workflows with scheduled jobs, supporting data ingestion pipelines, and backend logic organized around repeatable batch processing.

Result: Reduced payout processing turnaround from 5 days to 1 day and handles ~300-400 monthly disbursements.

Open Source

Public repos showing the actual code behind the work above

Three small, fully tested projects — each with CI running on every push — that mirror real patterns from production systems: secure API design, safe LLM agent permissions, and reliable ETL processing.

API & Security

Secure Payroll Disbursement API

A FastAPI microservice demonstrating production payroll-system patterns: OAuth2 client-credentials auth, decimal-correct financial math, and hybrid AES/RSA encryption for payment payloads.

  • Decimal-safe rounding, never bare floats, for currency math
  • Fresh AES-256 key per request, RSA-encrypted, for each payload
  • Batched disbursement with per-record success/failure tracking
FastAPI OAuth2 AES/RSA pytest

10 tests passing · CI on every push

Payroll API request flow A request authenticates via OAuth2, computes net pay with decimal-safe math, encrypts the payload with hybrid AES and RSA encryption, then disburses it in a tracked batch. 1 OAuth2 Auth client credentials 2 Compute Pay decimal-safe math 3 Encrypt AES-256 + RSA hybrid 4 Disburse batched, tracked
View Code

LLM Agent Safety

Safe LLM Agent — Tiered Permissions

A demo of the safety pattern behind a production Teams-based agent: the model can propose a mutation or a payment, but a human must approve it before it executes.

  • Read actions execute immediately; mutation and money actions never do
  • Approvals persisted to disk before execution, so a restart can't replay one
  • Money-tier actions serialized behind a lock to prevent concurrent payments
Anthropic API Agent Safety Python pytest

6 tests passing · CI on every push

Tiered agent approval flow The model proposes a tool call, which is checked against its tier. Mutation and money tier actions wait for a human to approve with a code, persisted to disk, before executing. 1 Model Call proposes a tool 2 Tier Check read, mutation, money 3 Human Approval code, persisted 4 Execute money tier locked
View Code

Data Engineering

Data Pipeline ETL

An ETL pipeline for messy batch data: strict per-row validation with specific rejection reasons, type-safe KPI aggregation, and batched writes with per-batch retry.

  • Every row gets a defined outcome: valid, or rejected with a reason
  • Type conversion happens explicitly during validation, never inferred
  • Batches tracked and retried independently; one failure never hides another
Pandas ETL Python pytest

25 tests passing · CI on every push

ETL pipeline flow Raw batch data is ingested, validated with per-row rejection reasons, aggregated into per-worker KPIs, and written out in batches with capped retries. 1 Ingest raw CSV batch 2 Validate typed, w/ reasons 3 Aggregate per-worker KPIs 4 Batch Write capped retries
View Code

Power Platform Work

Business applications and approval systems built for real internal teams

Internal Tools

Internal Automation Platform

Built and deployed 15+ Power Apps canvas applications, including 4 flagship systems: an exception-handling app, an Employee Self-Service app for attendance and designation tracking, an HR Headcount app for the MIS team, and an IT Audit Records app for the IT team. These apps were integrated with Power Automate workflows to reduce repetitive work while giving teams a more consistent process layer for everyday operations.

  • Dataverse-backed application flows for business data
  • Workflow orchestration through Power Automate
  • Operational dashboards and automated notifications

Approvals

Team Approval Automation System

Created a multi-level approval system using Power Apps, Power Automate, and Dataverse. The main goal was to make approvals visible, auditable, and less dependent on informal coordination through email or chat.

  • Multi-step approval routing with status tracking
  • Automated notifications at each stage of the process
  • Dataverse as the central workflow data layer

Document Flow

Document Management System

Built a Power Apps based document workflow connected to SharePoint and Microsoft Graph. The application supported document tracking, reporting, and process visibility for teams handling internal records.

  • SharePoint-integrated document handling
  • Graph API support for reporting and automation
  • Clearer tracking across document lifecycle steps

Operations Workflow

Attendance Automation System

Created a Power Apps interface linked with monitoring APIs and backend automation steps to improve attendance operations. The solution connected user facing workflows with the data processing needed behind the scenes.

  • External monitoring API integration
  • Operational workflow support through automation
  • User interface for simpler attendance handling

Architecture

How I usually structure automation-heavy systems

Most of my projects follow the same engineering idea: keep the user-facing layer simple, move validation and orchestration into backend services, and make integrations explicit so workflows stay observable instead of becoming hidden operational logic.

That typically means combining an interface layer such as Power Apps or a web frontend with API services, scheduled jobs, workflow orchestration, and a storage layer like Dataverse or SharePoint. External systems are treated as inputs that need validation, normalization, and traceable processing.

Interface Layer

Forms, internal apps, or frontends that collect inputs and show status.

Service Layer

APIs handle validation, business rules, orchestration, and integrations.

Automation Layer

Scheduled jobs and Power Automate flows keep recurring work consistent.

Data Layer

Dataverse, SharePoint, and structured datasets store reliable outputs.

Payroll to disbursement request flow An HR trigger flows through Power Automate, a FastAPI payroll engine, Dataverse, an encrypted disbursement step, and an async callback, with a return path notifying HR when complete. Power Automate trigger compute + batch write encrypt + disburse (ICICI) status update + notify HR 1 HR Trigger Power Apps 2 Power Automate Dataverse event flow 3 Payroll Engine FastAPI, KPI compute 4 Dataverse OData $batch write 5 Disbursement AES/RSA to ICICI 6 Reverse-Feed API async status callback

FAQ

Quick answers about the kind of work I do

What kind of problems do I solve?

I work on internal systems where business processes have outgrown manual coordination. That includes approvals, payroll support workflows, attendance processing, operational reporting, and integrations between business tools.

Where do I add the most value?

I add the most value where backend logic and business workflow design need to work together. I am comfortable translating operational pain points into APIs, automation rules, and maintainable process flows.

What technologies do I use most often?

Python, FastAPI, Flask, Pandas, Power Apps, Power Automate, Dataverse, SharePoint, and REST based integrations form the core of my day-to-day work.

What am I aiming for in a role?

I am most aligned with roles that involve backend systems, automation engineering, internal platforms, workflow orchestration, and business process tooling with real operational impact.

Contact

Interested in backend systems, internal tools, or automation work?

If you are building software for operations, workflow-heavy teams, or internal platforms, I would be glad to connect. This portfolio is focused on the kind of engineering work where reliable processes matter as much as clean code.

Email GitHub