I-983 Section 5 範例：STEM OPT 訓練計畫填寫參考（史蒂文斯2026）

為什麼 Section 5 是 I-983 的審核重點

I-983（Training Plan for STEM OPT Students）由學生與雇主填寫後，提交給學校 DSO（Stevens ISSS）審核、由其簽發 STEM OPT I-20，並非提交給 USCIS。Section 5 是 DSO 審核與日後 SEVP site visit 的重點，必須具體、詳細、有數據或時間表，清楚證明工作內容與你的 STEM 學位直接相關。

以下提供 Computer Science／Software Engineering 與 Financial Engineering／Business Analytics 兩個常見範例，供史蒂文斯理工學院（Stevens）研究生填寫時參考。範例僅作格式與寫法示意，實際內容須依你的真實職位、雇主與課程如實填寫。

範例一：Computer Science / Software Engineering

Goals and Objectives:

• Goal 1: Master full-stack development using modern frameworks (React, Node.js, AWS) to improve system performance and scalability.
• Goal 2: Apply machine learning techniques to develop intelligent features, such as recommendation engines or predictive analytics models.
• Goal 3: Gain expertise in DevOps practices, CI/CD pipelines, and cloud infrastructure management to support enterprise-level deployments.

How Goals Will Be Achieved:

• Goal 1: In the first 6 months, the student will participate in sprint planning and contribute to 3–4 major feature developments, guided by senior engineers (bi-weekly code reviews).
• Goal 2: During months 7–12, the student will work on an ML project using Python/TensorFlow, completing model training and deployment by Q3 2027.
• Goal 3: The student will complete internal AWS certification training and lead one infrastructure automation project within the first 12 months.

Employer Oversight and Supervision: The student will report directly to the Senior Software Engineering Manager [Name]. Supervision includes weekly 1:1 meetings, quarterly performance reviews, and access to the internal mentorship program. All code will undergo peer review before deployment.

Performance Evaluation Process: Formal evaluations will occur at 12 months and 24 months using the company's standard performance review system, measuring technical contribution, code quality, and project delivery against the above goals.

範例二：Financial Engineering / Business Analytics

Goals and Objectives:

• Goal 1: Build and validate pricing models for derivatives using stochastic processes and Monte Carlo simulations.
• Goal 2: Develop machine learning-based credit risk prediction models to support real-time decision-making.
• Goal 3: Optimize trading strategies through back-testing and high-frequency data analysis.

How Goals Will Be Achieved:

• Goal 1: Complete model development and validation for at least two derivative products within the first 6 months, with regular feedback from the Quantitative Research team.
• Goal 2: In months 7–18, implement ML models in Python/R and integrate them into production risk systems.
• Goal 3: Conduct monthly strategy performance reviews and achieve at least 10% improvement in Sharpe ratio for assigned portfolios by the end of Year 1.

Employer Oversight and Supervision: Daily supervision by the Head of Quantitative Research. The student will join daily stand-up meetings, weekly model validation sessions, and have access to proprietary datasets and computing resources. Progress will be tracked via Jira and internal dashboards.

Performance Evaluation Process: Evaluations at 12 and 24 months will assess model accuracy, business impact, and adherence to regulatory standards, documented through formal written reports signed by the supervisor.

填寫小提醒（Stevens 學生適用）

• 越具體越好：使用課程名稱（如 CS 600 Algorithms）、工具（Python、AWS、MATLAB）與可衡量的預期成果（如 improve accuracy by 15%）。
• 長度：每個欄位建議 4–8 句，必要時可附加額外頁面（Addendum）。
• 與雇主合作：讓雇主（Supervisor／HR）共同審核與簽名。
• Stevens 專屬資源：使用學校提供的 2026 更新版 I-983 Guide，並先把草稿上傳給 ISSS 審核。

官方資源與重要提醒

官方資源：DHS I-983 表格與說明（ice.gov）、Study in the States 的 STEM OPT Hub。填寫前後都應對照官方最新版本。

本文僅為格式與寫法的教育性參考，不構成移民、法律或個案建議。I-983 提交給學校 DSO（Stevens ISSS）而非 USCIS；STEM OPT、I-983 與 OPT Extension 的資格、時限與填報要求依美國移民法規與你的個案而定，草稿完成後務必交由 Stevens ISSS 或合資格的移民專業人士最終審核，避免 DSO 退件或日後查核問題。