AH.
AH.

TAKE A LOOK AROUND

I BUILD SOFTWARE THAT MATTERS.

Software engineer working across backend systems, applied AI, and full-stack products. Turning complex problems into reliable, useful software.

SCROLL TO FOLLOW THE THREAD

01/ REAL-WORLD ENGINEERING

BEYOND
THE CODE.

Engineering is not only about building features. It is about understanding failures, connecting systems, and making software dependable.

CURRENT CHAPTER / ISLAMABAD, PK

Target Systems/ OOMI CRM

Junior Software Engineer

JULY 2025 — PRESENT
The full experience
01

Production issue investigation

Follow a reported failure through OOMI CRM, trace the affected records, and work with the engineering team to establish the root cause.

OOMI CRM · Jira Service Management
02

Backend & database operations

Investigate SQL Server relationships and application inconsistencies. Apply controlled data corrections and validate the resulting business workflow.

SQL Server · SSMS · Data integrity
03

APIs & connected systems

Trace REST requests, JSON payloads, OAuth, and authentication issues across integrations including Xero and GoCardless.

REST APIs · OAuth · Xero · GoCardless
04

Workflows & application behavior

Debug CRM processes, workflow automation, and frontend behavior. Connect what a user sees to the application and data underneath.

CRM workflows · Browser developer tools
05

Monitoring & root cause analysis

Use AWS CloudWatch and operational logs to investigate errors, build context around a failure, and support application reliability.

AWS CloudWatch · Logs · Troubleshooting
06

Engineering & client collaboration

Coordinate investigations with engineering colleagues and UK stakeholders. Document the findings, validate fixes, and prepare troubleshooting guidance.

Technical documentation · UK client communication
THE CHAPTERS BEFORE
Dec 2024 – Jul 2025

Software Engineer Intern

Target Systems

Jul 2024 – Sep 2024

Machine Learning Intern

Codic Solution

Jul 2023 – Aug 2024

Technical & Digital Operations

Pure Oxygen Therapy (UK)

02 / MY TOOLKIT

THE TOOLS
BEHIND THE WORK.

From backend APIs and databases to computer vision and intelligent applications.

LANGUAGES & DEVELOPMENT
  • Python
  • TypeScript
  • JavaScript
  • C#
  • React
  • Next.js
  • FastAPI
  • Django
  • Node.js
  • Tailwind CSS
AI, DATA & ENGINEERINGBUILT TO WORK TOGETHER
  • PostgreSQL
  • SQL Server
  • PyTorch
  • TensorFlow
  • scikit-learn
  • OpenCV
  • Docker
  • AWS
  • Git
  • Postman

03/ SELECTED WORK

All projects

PROBLEMS.
SYSTEMS.
SOLUTIONS.

01 / COMPUTER VISION / RETAIL

Turning camera feeds into practical retail intelligence through computer vision, analytics, and automated detection.

Retail Vista: a store seen by a cameraAn architectural store diagram with product shelves, a visitor, and a camera field of view. Video frames become the input to the vision pipeline.01 / THE STORECAMERA INPUTOPENCV / VIDEO FRAMESRetail Vista: detection with contextA sequence of drawn camera frames shows person detection and a candidate activity event. Dashed connectors distinguish frame processing from event analysis.02 / READING THE FRAMEFRAME / AOBJECT DETECTIONFRAME / BTEMPORAL CONTEXTDETECTION IS A SIGNAL. CONTEXT COMES NEXT.Retail Vista: a reviewable eventAn incident-review case file contains a camera frame, a temporal event strip, and a human-review annotation. It represents candidate events for inspection, not automatic proof of an incident.03 / A PERSON REVIEWSEVENT / REVIEW FILESAMPLECANDIDATE ACTIVITYFRAME SEQUENCE / AVAILABLE FOR INSPECTIONHUMAN REVIEWNOT A VERDICTREVIEWREQUIREDVISION → CONTEXT → A REVIEWABLE EVENT
01 Observe the store02 Understand the frame03 Put context in human hands

RETAIL VISTA

ILLUSTRATIVE PREVIEW / VISION WORKFLOW

A detection starts a review. Context makes it useful.

THE PROBLEM

Watching several camera feeds makes it difficult to spot and review potentially suspicious activity. A useful system needs to connect detections to context and a reviewable interface.

MY CONTRIBUTION

Dataset preparation, computer vision pipeline experimentation, application integration, and frontend/backend project work.

Explore case study

Python · YOLOv8 · OpenCV · PyTorch · Django · React · WebSockets

2ND POSITION / COMSATS OPEN HOUSE 2025

02 / RECOGNITION / ATTENDANCE

Connecting facial enrollment and recognition with real-time processing and structured attendance records.

VizonTrack: face enrollmentA line drawing of a face is surrounded by registration guides and five facial landmarks. Enrollment establishes the reference identity for later matching.IDENTITY ENROLLMENT01 / ENROLLMENTREFERENCEIDENTITYFACIALLANDMARKSILLUSTRATED FACE / NO BIOMETRIC DATAVizonTrack: features to identity matchingA second face is transformed into a feature vector. Parallel vector lines connect to a cylindrical similarity index, representing the intended InsightFace, ONNX Runtime, and pgvector recognition pipeline.RECOGNITION REPORT02 / FINDING A MATCHFEATURE VECTORSIMILARITYINDEXRECOGNITION AND ENROLLMENT STAY SEPARATEINSIGHTFACE / ONNX / PGVECTORVizonTrack: attendance as an eventAn open attendance ledger separates identity, event, and history. One sample check-in is recorded while an unknown identity remains explicitly unresolved.ATTENDANCE REGISTER03 / A RECORD WITH CONTEXTIDENTITY / SAMPLEATTENDANCE EVENTMEMBER ACHECK-INMEMBER BCHECK-INUNKNOWNNO MATCH / REVIEWENROLL → MATCH → INSPECTABLE HISTORY
01 Enroll an identity02 Match the face03 Keep a structured record

VIZON TRACK

ILLUSTRATIVE PREVIEW / CONCEPTUAL WORKFLOW

From a face in a frame to a record in a system.

THE PROBLEM

Attendance collection can become a separate manual process. This project explores how camera input, identity matching, and attendance history can fit into one application.

MY CONTRIBUTION

Independent application design, recognition pipeline exploration, backend/API development, and frontend workflows.

Explore case study

Python · FastAPI · InsightFace · ONNX Runtime · PostgreSQL · pgvector · Next.js · TypeScript · WebSockets

Independent project · In development

03 / DATA / VISUAL AUTOMATION

A visual workspace for building data transformation pipelines without manually chaining every operation.

DataFlowX: data starts with a schemaLoose CSV and JSON papers sit behind a tabular source sheet with named columns and illustrative rows. The known schema is the first input to a visual data workflow.01 / THE RAW MATERIALSOURCE / CSVDATASET / SAMPLEITEMREGIONVALUEANORTH24BWEST18CNORTH32DEAST—ROWS + SCHEMA / EXPLICIT INPUTSDataFlowX: a graph with an execution contractA branching pipeline routes a dataset through filter, transform, and group operations. Configuration, validation, and dependency connectors describe the work rather than decorative node positions.02 / MAKE THE LOGIC VISIBLE01 / FILTERvalue IS NOT NULL02 / TRANSFORMnormalize(value)03 / GROUPBY region / SUM valueVALIDATE FIRSTDEPENDENCY GRAPHCONFIGURE → VALIDATE → PROCESSDataFlowX: inspect the outputA folded result sheet shows a grouped dataset, a schema checklist, and an export symbol. Intermediate previews remain visible while sample data is processed.03 / UNDERSTAND THE RESULTOUTPUT / GROUPED DATAREGIONSUM / VALUENORTH56WEST18SAMPLE RESULTSCHEMA VALIDATEDSTEPS INSPECTABLEPREVIEW / NOT A PRODUCTION DATASETINTERMEDIATE STATE / A USEFUL OUTPUT
01 Start with a source02 Connect the operations03 Inspect the result

DATAFLOWX

ILLUSTRATIVE PREVIEW / CONCEPTUAL WORKFLOW

Make the transformation as clear as the result.

THE PROBLEM

Small data transformations often turn into hard-to-follow scripts. A visual workflow should make the sequence, configuration, and output of each step easier to understand.

MY CONTRIBUTION

Independent graph interface, frontend state management, processing contracts, and Python data-operation exploration.

Explore case study

Python · FastAPI · Pandas · React · TypeScript · React Flow · PostgreSQL

Independent project · In development

04 / REAL-TIME / COLLABORATION

A team collaboration platform built around channels, messaging, presence, and shared workspaces.

Connect: a workspace built around peopleThree illustrated team members connect to a shared workspace and its project and general channels. A workspace boundary defines membership.WORKSPACE MEMBERS01 / A PLACE FOR THE TEAMMEMBER AMEMBER BMEMBER CSHARED WORKSPACEMEMBERSHIP / ACCESS# GENERAL# PROJECTS# ENGINEERINGCHANNELS KEEP A CONVERSATION IN CONTEXTConnect: a conversation becomes workA drawn conversation sheet shows a project discussion leading to an attached task. The message and task retain a shared project context.PROJECT CONVERSATION02 / KEEP THE CONTEXT# PROJECT / CONVERSATIONAMEMBER AReady for a review?BMEMBER BLinking the task here.PROJECT TASK / SAMPLEReview the workflowDISCUSSION → A PIECE OF WORKILLUSTRATIVE CONVERSATION / FICTIONAL MEMBERSConnect: shared state and presenceA cloud-like collection of connected message, task, and membership documents represents the shared relational model, with explicit presence indicators around its edge.SHARED WORK03 / CONNECT THE PIECESMESSAGESPROJECT TASKSONLINEONLINEAWAYMESSAGES / TASKS / MEMBERSHIP / PRESENCE
01 Find your workspace02 Keep the conversation03 Move the work forward

CONNECT

ILLUSTRATIVE PREVIEW / CONCEPTUAL WORKFLOW

People, conversations, and work. In the same place.

THE PROBLEM

Team discussions lose context when conversations and tasks live in separate places. Connect explores a workspace where a discussion can lead directly to a piece of work.

MY CONTRIBUTION

Independent workspace interface, application state, relational project models, and real-time architecture exploration.

Explore case study

Next.js · TypeScript · Node.js · PostgreSQL · WebSockets

Independent project · In development

MORE LINES OF INQUIRY05

Man Overboard Detection

COMPUTER VISION / MARITIME SAFETY06

Multiple Disease Prediction

MACHINE LEARNING / ACADEMIC PROJECT

04/ THE JOURNEY

FROM CURIOSITY
TO ENGINEERING.

2023
Foundations

Learning the underlying logic.

Software engineering at COMSATS: programming, algorithms, databases, and the fundamentals behind dependable applications.

University and academic software projects.
2024
Machine learning & vision

From images to understanding.

At Codic Solution, I worked with YOLOv8, OpenCV, dataset preparation, model experimentation, training, and evaluation.

Machine Learning Intern · July–September 2024.
2025
Professional engineering

Working inside real systems.

Graduated from COMSATS and continued at Target Systems / OOMI, investigating CRM issues, tracing SQL records, troubleshooting integrations, and contributing React/.NET functionality.

Junior Software Engineer · July 2025–present.
2026
Building complete systems

Connecting more of the pieces.

Independent products bring together Python services, data processing, full-stack interfaces, and applied AI. These are development projects, with capabilities being built and explored.

Independent work and ongoing experimentation.

05 / THE PERSON BEHIND THE SYSTEMS

EVERY SYSTEMCARRIES AHUMAN IMPRINT.

Behind the data, the APIs, and the decisions — a person who likes understanding how things work. And figuring out why they don’t.

My path started at COMSATS, through experiments in computer vision and into production engineering at Target Systems. Today, I work across backend systems, applied AI, and full-stack products.

Anish HamayoonSOFTWARE ENGINEER / BACKEND & APPLIED AIISLAMABAD, PAKISTAN
Resume More about me

06 / LET’S CONNECT

HAVE APROBLEMWORTHUNTANGLING?

A useful product. A complex system. A problem that needs a fresh pair of eyes.

anishhamayoon06@gmail.com
GITHUB LINKEDIN RESUME
GOOD WORK STARTS WITH A CONVERSATION.