It’s been exactly a year since I started my new position as AI, ML, NLP Engineer and Data Scientist, so now is the time to reflect. Technology-wise it is an ideal moment for introspection and planning, particularly in the field of NLP, which has seen transformative changes with the rise of Large Language Models (LLMs). In times of great change and opportunity it is important to make time to stand still and reflect, here’s how and why I do this every year:

Reflection

The Significance of Annual Career Reflection

Adapting to Revolutionary Changes

The emergence of LLMs has radically altered our landscape. Reflecting annually helps in adapting to these shifts, ensuring we remain at the forefront of continuous learning and innovation.

Goal Setting: A Strategy for Advancement

Carefully evaluating the past year’s goals and achievements isn’t just about self-improvement; it’s crucial for career progression in a competitive field like NLP.

Celebrating Achievements

Documenting every success, no matter the scale, is vital. It’s not just a record of progress; it’s a testament to your capabilities and contributions, essential for career advancement. Any Data science, AI and NLP engineering project will have challenges to overcome, some projects will feel like a continual series of issues. Knock one down another pops up.

To maintain sanity it is important to celebrate success and highlight wins.

Career Review Best Practices for NLP Engineers

Continuous Learning in the Era of LLMs: Tech and the human element

Staying abreast of advancements, particularly in LLMs, is essential. This continuous learning is what keeps us relevant and innovative in NLP.

Prompt Engineering: Crafting precise prompts can significantly boost LLM performance in zero and few-shot tasks. Like finding the right words in a magic spell, I doubled a pipeline’s performance by refining prompts alone (Brown et al., 2020).
Open Source LLMs: These models open doors to customization, allowing for fine-tuning, self-hosting for quicker inference, and enhancing data privacy. They’re not just open books; they’re whole libraries waiting to be explored (Wang et al., 2019).
Retriever Augmented Generation (RAG): A revolutionary approach in QA, RAG combines the depth of knowledge retrieval with LLMs’ generative capabilities. While it promises to be LLMs’ killer app, turning this potential into production reality is like herding digital cats—challenging but not impossible (Lewis et al., 2020).

On the human side one of the largest tasks will be

Communication and expectation management: Sparked by OpenAI’s releases of ChatGPT and GPT-4, expectations around AI capabilities have skyrocketed. We have seen a surge in interest, demand and investment in our field. This excitement, however, brings the challenge of managing expectations realistically without dampening enthusiasm. As NLP engineers, our role extends beyond development; we become part educators, tasked with clarifying AI’s current strengths and limitations. This involves navigating ambitious assumptions with clear communication, ensuring stakeholders understand what’s achievable versus what remains aspirational. It’s about refining ambition with (budgetary) reality, guiding projects toward focused, achievable goals, and steering the conversation towards a future where AI’s potential is both understood and maximized.

Networking: Collaboration and Idea Exchange

Building connections is about more than just opportunities. It’s about being part of the evolving conversation in NLP, sharing insights, and collaborating on groundbreaking projects.

Conducting an Effective Year-End Review

A well-executed review can be a game-changer. Here’s how:

Reflect on Past Goals: Analyze outcomes, lessons learned, and strategies used.
List Achievements: Every project, paper, or breakthrough counts.
Gather Feedback: Seek insights from peers, mentors, and the community.
Set Future Objectives: Make them precise, meaningful, and aligned with industry trends.
Plan for Development: Focus on skills, collaborations, and staying updated with LLM advancements.

Incorporating the Career Goals Template

To structure this process, I’ve created a detailed template that you can use to guide your reflection and planning. It covers everything from past reflections to future goals and professional development plans. (Template to be added)

Happy reflecting!

The Review Template

# Career End-of-Year Review

## Executive Summary
- Brief overview of key achievements
- Summary of goals met

## Reflection on Past Goals
- **Goal 1:**
  - Was it achieved? 
  - Challenges faced:
  - Lessons learned:
- **Goal 2:**
  - Was it achieved?
  - Challenges faced:
  - Lessons learned:
- *(Continue for other goals)*

## Achievements and Wins
- **Achievement 1:**
  - Description:
  - Impact on the team/organization:
- **Achievement 2:**
  - Description:
  - Impact on the team/organization:
- *(Continue for other achievements)*

## Quantified Achievements
- **Achievement/Data Point 1:**
  - Description:
  - Quantitative Measure:
- **Achievement/Data Point 2:**
  - Description:
  - Quantitative Measure:
- *(Continue for other quantified achievements)*

## Feedback Received
- **Feedback from [Manager/Peer/Client]:**
  - Key points:
  - Areas of improvement:
- *(Continue for other feedback received)*

## Areas for Improvement
- **Area 1:**
  - Description:
  - Plan for improvement:
- **Area 2:**
  - Description:
  - Plan for improvement:
- *(Continue for other areas)*

## Goals for Next Year
- **Goal 1:**
  - Description:
  - Action Plan:
- **Goal 2:**
  - Description:
  - Action Plan:
- *(Continue for other goals)*

## Professional Development Plan
- **Skill/Qualification 1:**
  - Plan to achieve/learn:
- **Skill/Qualification 2:**
  - Plan to achieve/learn:
- *(Continue for other skills/qualifications)*

## Conclusion
- Final thoughts
- Readiness for the coming year

## Appendices (Optional)
- Supporting documents
- Additional data or reports

Celebration

My 2023 Achievements

I am not going to share the challenges, learning points because that’s too personal. But have my achievement list:

Achievement 1: Data Scientist and NLP Engineer Role
- Held a position focusing on Data Science and Natural Language Processing within the data engineering and science department of a major corporation.
- Successfully navigated a competitive job market, resulting in multiple job offers through a strategic analysis of opportunities.
- Impact: Achieved a role with an excellent balance between work demands and personal life, focusing on NLP and Machine Learning. Contributed to significant improvements in project management processes and inter-departmental collaboration, enhancing organizational efficiency and client communication.
Achievement 2: Independent Freelance NLP Engineer
- Founded a consulting business to provide expertise in Machine Learning and Natural Language Processing.
- Collaborated on advanced ML/NLP projects with a technology startup, contributing to innovative solutions.
- Impact: Delivered high-value consulting services, focusing on the development of complex NLP/ML models and pipelines, resulting in highly engaging and financially rewarding projects.

Professional Experience at a Leading Organization (Achieved at 80% Employment):

Achievement 1: Automation Initiative for Customer Service
- Led the enhancement of an automated system for classifying customer service inquiries, improving efficiency and accuracy.
- Spearheaded the upgrade of the system using advanced machine learning techniques, enabling faster processing and better resource allocation.
- Conducted comprehensive analysis to optimize the system, resulting in a well-received prototype demonstrating significant business value.
Achievement 2: Strategic Project Planning and Funding Acquisition
- Initiated and led the strategic planning of several AI and data science projects aimed at achieving long-term organizational goals.
- Played a key role in securing substantial funding for the department by developing detailed project plans and demonstrating potential business impacts.
Achievement 3: Virtual Assistant Development
- Developed a prototype for an internal virtual assistant using emerging technologies, designed to improve customer service efficiency.
- Conducted research and development activities to explore new methods for document analysis and interaction, culminating in a successful demonstration of the prototype’s capabilities and securing project funding.
Achievement 4: Feasibility Study for Automated Mail Classification
- Conducted a feasibility study on automating mail classification for a specific department, addressing unique data challenges.
- The study provided valuable insights, leading to informed decisions regarding the project’s potential and strategic direction.
Achievement 5: Transition of Mail Automation from Prototype to MVP
- Managed the transition of an automated mail classification system from a prototype to a minimum viable product (MVP), overseeing project planning and technical development.
- Contributed to software engineering efforts, enhancing system capabilities and achieving a successful launch of the MVP.
Achievement 6: Development of an Internal Virtual Assistant
- Led the development of an internal virtual assistant aimed at improving customer service operations, from concept to MVP.
- Oversaw data collection and system improvement efforts, resulting in an ongoing successful MVP release.

Freelance Contributions to Sentometrics:

Achievement 1: Development of an NER System for News Monitoring
- Developed a system for automatically identifying company and personal names in news articles, enhancing the metadata quality for a news monitoring product.
Achievement 2: Expansion of Company Lexicon
- Improved the coverage of a lexicon used for economic news monitoring, leveraging multilingual resources and similarity algorithms to enhance product capabilities.
Achievement 3: Knowledge Graph Project for ESG Monitoring
- Played a pivotal role in securing funding for a project aimed at developing a knowledge graph for Environmental, Social, and Governance (ESG) monitoring, contributing to the technical and strategic planning.
Achievement 4: Improvement of ESG Materiality Assessment
- Led the enhancement of an assessment system for evaluating ESG materiality, significantly improving the accuracy and efficiency of the process through iterative development and testing.

References:

Brown, T. B., et al. (2020). Language Models are Few-Shot Learners. arXiv preprint arXiv:2005.14165.
Wang, A., et al. (2019). GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding. arXiv preprint arXiv:1804.07461.
Lewis, P., et al. (2020). Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks. arXiv preprint arXiv:2005.11401.