Ali Astra

AI Engineer with 2 years of experience specializing in Deep Learning, Computer Vision, and edge deployment. Experienced in designing, training, and optimizing deep learning models for real-time and edge-based vision systems, with a strong focus on performance, efficiency, and scalability. Proven ability to develop and deploy production-ready solutions, including real-time inference, API integration, and end-to-end machine learning pipelines.

Proficient in Python with hands-on experience in TensorFlow and PyTorch, and supported by strong knowledge of FastAPI, OpenCV, Scikit-learn, ONNX, and object detection frameworks such as YOLO. Additionally, possesses working knowledge of C and C++ to support optimization and deployment on edge devices. Familiar with Google Cloud Platform (GCP) and MLOps practices and containerization using Docker and Gaps-based deployment to ensure reliable, maintainable, and scalable AI solutions for real-world applications.

Full Stack AI Engineer - Indonesia AI Institute - Bandung, West Java

(2026-01)

• Developed and maintained an AI Learning Platform with a scalable backend architecture, supporting chat systems, analytics, course management, and task orchestration, improving system modularity and development efficiency.
• Built and integrated LLM-powered chat systems with session management and chat history pipelines, enabling personalized learning experiences and improving user engagement.
• Designed and deployed a mental health chatbot used by 100+ users, implementing structured system prompting and template-based prompt engineering to improve response relevance and consistency.
• Implemented real-time response streaming using Server-Sent Events (SSE), reducing perceived latency and improving response delivery time by up to 30–40%.
• Applied guardrails and output constraints to enhance safety and reliability of LLM responses, reducing invalid or unsafe outputs in sensitive use cases.
• Developed and deployed end-to-end AI systems with API integration and production-ready pipelines, ensuring scalability and maintainability across services.
• Utilized Docker and MLOps practices to streamline deployment, improving development-to-deployment cycle efficiency and system reliability.

Educator (Coding & AI) - Yayasan Sakata Innovation Center - Tasikmalaya, West Java

(2025-07 - 2025-12)

• Responsible for delivering instructional materials and hands-on training on Machine Learning, Deep Learning, and the use of Large Language Models (LLMs) to 60+ senior high school and vocational school teachers (SMA/SMK) in Tasikmalaya and Ciamis.
• Developed practical ML notebooks and experiments using TensorFlow and Hugging Face APIs.

Machine Learning Mentor - Codingcamp by DBS Foundation - Remote

(2025-02 - 2025-05)

• Mentored 25 learners in machine learning concepts, Python, and applied ML projects.
• Improved course completion rate to 70% through structured mentoring and performance tracking.

Assistant Lecturer for Database - Universitas Siliwangi - Tasikmalaya, West Java

(2025-01 - 2025-05)

• Supported and guided 90+ undergraduate students across multiple classes during laboratory and tutorial sessions.
• Taught relational database design, normalization, SQL querying and assisted students with database implementation.

Computer Vision Engineer - Siliwangi Vision - Tasikmalaya, West Java

(2024-08 - 2024-11)

• Reduced manual monitoring effort by up to 80% by developing a real-time people and vehicle counting system with 92.1%+ detection accuracy using optimized computer vision models.
• Achieved sub-200ms inference latency and stable real-time performance by optimizing models for edge deployment and efficient video stream processing.
• Improved counting accuracy in complex environments by implementing polygon-based ROI, enabling flexible adaptation to various camera perspectives.
• Delivered actionable traffic insights by classifying vehicles (cars, trucks, buses) and processing thousands of events daily into time-series analytics.
• Reduced attendance fraud and improved verification reliability by implementing face detection with liveness detection, achieving >95% spoof detection accuracy.

Bangkit Academy 2023 Machine Learning Engineer (Apprenticeship) - AMATI Indonesia - Jakarta (Remote)

(2023-09 - 2023-11)

• Developed and deployed a machine learning–based product recommendation system for a supply chain application, collaborating with cross-functional teams (Cloud and Android).
• Built end-to-end ML pipeline including data cleaning, feature engineering, vectorization, and model evaluation for a production recommendation system.
• Implemented a content-based recommendation system using QR-code-based features, achieving 95% accuracy and integrating the model via Flask-based REST APIs.
• Built and deployed RESTful ML inference APIs using Flask, enabling real-time model serving and system integration.

Bachelor Degree of Informatics - Informatics - Universitas Siliwangi (2021-08 - 2025-12)