Skip to content
View Danimannnm's full-sized avatar

Highlights

  • Pro

Block or report Danimannnm

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Danimannnm/README.md
Adnan Khaliq — Cyber Neon Pixel Banner

LinkedIn · Email · GitHub

AI / ML ENGINEER

SOFTWARE ENGINEER AI / ML PYTORCH DATA PIPELINES CLOUD ML

ABOUT

I build applied AI and machine learning systems — from raw data and feature engineering to trained models, evaluation, and deployment-ready prototypes.

My experience spans computer vision, forecasting, data analysis, and full ML pipelines, with a strong emphasis on writing code that is measurable, debuggable, and useful beyond notebooks.

I care about clean data flows, correct experimental design, and turning models into tools people can actually interact with.

TOP FOCUS

AREA WHAT I WORK ON WHY IT MATTERS
Applied AI / ML Model training, evaluation, experimentation Correct decisions need reliable models
Data Pipelines Ingestion, preprocessing, feature engineering Garbage in = garbage out
Computer Vision Detection, segmentation, spatial & image-based models High-impact real-world applications
Forecasting & Time Series Time-aware modeling, indicators, evaluation Most demos get this wrong
Tooling & Visualization Dashboards, analysis tools, model inspection Visibility beats blind accuracy
Systems & Integration APIs, auth, cloud services, deployment Models must live in systems

TOOLBOX

AI / ML
PyTorch · TensorFlow/Keras · Deep Learning · Model Evaluation · Feature Engineering

DATA & ANALYSIS
Pandas · NumPy · Time Series · Visualization · APIs / Scraping

BACKEND / FULL-STACK
Node.js · Express · React · REST APIs · Auth (JWT / OAuth)

CLOUD / DEVOPS
Azure · AWS · basic GCP · Docker · Kubernetes · CI/CD exposure

DATABASES
MongoDB · MySQL · SQL Server

WHAT I’M LOOKING FOR

AI / ML internships or early-career roles where I can work on real datasets, end-to-end pipelines, and applied machine learning problems — especially in environments that value correct evaluation and production-minded thinking.

FEATURED PROJECTS

[AI / ML] END-TO-END ML PIPELINES & EXPERIMENTATION

WHAT IT DOES
Designs and evaluates machine learning pipelines covering data ingestion, preprocessing, training, and evaluation.

TECH STACK
Python · PyTorch · Feature Engineering · Evaluation Metrics

WHY IT’S INTERESTING
Most ML issues come from pipeline mistakes, not model choice.

OUTCOME
Reproducible experiments with clear performance comparisons and reliable results.

[CV] INDUSTRIAL OBJECT DETECTION — CLOUD BENCHMARKING

WHAT IT DOES
Benchmarks cloud-based object detection systems using a shared industrial dataset.

TECH STACK
Azure Custom Vision · Google Vertex AI · Roboflow · Python · Dashboard UI

WHY IT’S INTERESTING
Real deployments care about iteration speed, dataset handling, and cost — not just accuracy.

OUTCOME
Identified the most practical platform based on accuracy, efficiency, and workflow fit.

[ML] NEURAL FORECASTING & TIME-SERIES ANALYSIS

WHAT IT DOES
Forecasts next-day stock prices using neural models with proper time-aware validation.

TECH STACK
Python · Neural Networks · Technical Indicators · Streamlit

WHY IT’S INTERESTING
Many forecasting demos leak future data — this pipeline avoids that entirely.

OUTCOME
Interactive dashboard comparing models, error metrics, and trading simulations.

[FULL-STACK / WEB] APMC JUDGING PLATFORM

WHAT IT DOES
A complete web-based judging platform where users can submit entries, reviewers can evaluate them, and results are processed and displayed through a structured workflow.

TECH STACK
React · Node.js · Backend APIs · Database · Auth · System Design

WHY IT’S INTERESTING
Judging platforms are state-heavy systems — handling submissions, roles, evaluation logic, and result consistency requires careful backend design, not just UI work.

OUTCOME
Fully functional end-to-end system with a clean frontend, reliable backend logic, and well-defined judging workflows.

[FULL-STACK] PURRRFECT MATCH — SYSTEM DESIGN & INTEGRATION

WHAT IT DOES
Full-stack web platform with authentication, media handling, and backend services.

TECH STACK
MongoDB · Express · React · Node.js · OAuth · JWT · Cloudinary · PostMark

WHY IT’S INTERESTING
Demonstrates how ML systems eventually fit into real applications.

OUTCOME
Production-style app with clean UX and robust backend integration.

CONTACT ME

If you’re working on AI, machine learning, data, or applied systems, I’m always open to conversations — internships, collaborations, or technical discussions.

Email LinkedIn GitHub

Mian Adnan Khaliq

Pinned Loading

  1. 3D-PointCloud-Segmentation-OBB 3D-PointCloud-Segmentation-OBB Public

    Proof-of-concept for 3D pole detection and dimensioning using Toronto-3D LiDAR data. Trains a KPConv segmentation model, extracts pole masks, and applies clustering with bounding boxes or cylinder …

    Python