Skip to content
DSML.KZ
AI

Alan Igilikov

@olanmeister

CV engineer

I am CV engineer. Strong background in python, ml frameworks, gstreamer, deepstream, devops/mlops, backend, db management, ml research and communication.

Career

Experience: 2 years

Experience

Middle Computer Vision Engineer

Sergek Group

Oct 2025 — Present

Builds production video analytics, including seatbelt detection, face analysis, liveness checks, and face tracking improvements.

• Built a two-stage hybrid pipeline for driver seatbelt detection: YOLO-World and SAM3 localise the driver region in real time and filter frames at the edge, then verified triggers go to a Qwen-VL vision-language model served on vLLM, whose visual reasoning keeps false positives low. • Designed and shipped a multi-task face analysis model: six attribute heads (age, gender, emotion, beard, glasses, mask) trained on shared ArcFace features and merged with the backbone into a single ONNX graph, so one forward pass returns the identity embedding and all attributes.

Junior Software Engineer

TargetAI Limited

Apr 2025 — Sep 2025

Co-developed an internal support-team tool for managing local product usage.

• Co-developed “Backyard”, an internal tool for the support team to manage local product usage.

Education

BSc in Computer Science

Nazarbayev University

2021 — 2025

Selected projects

Item Sales Prediction: Classical ML vs. Recurrent Networks

Machine Learning Group Project at Nazarbayev University (Aug–Dec 2024). Compared classical models (XGBoost, Seasonal ARIMA) with an LSTM-based RNN on a Kaggle sales case; engineered seasonal and numeric features and tuned hyperparameters, dropout and optimiser settings. Best result: F1 score of 0.98 with RNN + LSTM.

Electric Motor Fault Detection from Vibration Signatures

Deep Learning Group Project at Nazarbayev University (Aug–Dec 2024). Automated extraction of raw motor vibration data into a large labelled dataset and trained a TCN model to predict motor faults (RAE 28%).

Skills

PythonC++SQLJavaScriptPyTorchTensorFlowscikit-learnXGBoostNumPypandasCNNRNN/LSTMTCNFace recognition (ArcFace)Anti-spoofingAttribute classificationObject detection (YOLO-World)Segmentation (SAM3)Vision-language models (Qwen-VL)Tracking and clustering