Manas Dasgupta

Few additional points: Timebox Experimentation: Allocate a fixed time for experimentations, such as trying different models or feature engineering techniques. Timeboxing prevents endless tweaking and ensures that the project progresses within the set timeline. Risk Identification and Mitigation: Identify potential risks early on, such as data quality issues, lack of stakeholder buy-in, or deployment challenges. Develop contingency plans for these risks to avoid delays. Leverage Prototypes and MVPs: Develop prototypes or Minimum Viable Products (MVPs) early in the project to validate key assumptions and gather feedback. This reduces the risk of investing time in a full solution that may not meet the requirements.

Its a highly achievable and structured process: - You need to master a language like Python (which has a small learning curve) - Understand the basics of AI / Data Science. - Start learning Classical Algorithms: Supervised and Unsupervised Learning, Regression and Classification methods. - Hands on: get into a platform like Kaggle or Github and seek out Projects to solve on a regular basis. Build your portfolio on these platforms. - Understand experimentation, evaluation and optimization techniques. - Learn Neural Networks and based on your interest, branch out to some specialization like Image Processing, Natural Language Processing. - Gen AI techniques now are a very good addition to your AI skills. Good luck!

1. Undertake an exploration of finding useful Business Cases that with an AI Solution will achieve tangible and significant benefits. 2. Develop AI Capability with a long term view. AI is diverse - so, develop diverse AI capability - this may include Classical Algorithms, Neural Nets (Computer Vision, Natural Language Processing), Generative AI and so on. 3. Take up some pilots with low hanging cases. 4. Establish a solid benefits measurement mechanism. 5. Create a culture of experimentation. 6. Benchmark your organization on AI Capability to the Industry. 7. Bring Business and Technology stakeholders together in the Journey. 8. Publish and celebrate your successes.

The most important component is to identify the business case and desired benefits. On the Tech part, make sure you identify the right Model for your business case otherwise the outcome may not be adequate, if it necessitates, provision for Fine Tuning of the model with data related to your context. Optimized Prompts are critical to extracting the right information. Choosing the right Prompt Type and Agent Types are important for an optimized execution. Look for any specialized External API that may augment the outcome or speed up the process.

This is a skill that one needs to strengthen more and more as we grow in the ladder: Be aware of your own biases and vulnerabilities as people would try to exploit these. Identifying your own weaknesses can help you be more vigilant. Establish upfront what you are and are not willing to discuss or negotiate. This helps prevent the conversation from veering off course. Pay attention to the person's body language and tone of voice. These can sometimes reveal their true intentions even if their words don't. Don't be afraid to disagree with someone, even if they are in a position of authority. Express your opinions and concerns clearly. If you're unsure about something, consult with a trusted colleague or advisor before making a decision.

Storytelling can help resolve conflicts. My two cents: Select a story that is relevant, relatable, and emotionally resonant to your audience. Highlight shared values: Use stories that showcase values you share with your stakeholders. Emphasize common ground for collaboration. Illustrating consequences: Show the potential outcomes of different choices, the benefits of finding common solutions. Building Empathy: Stories connect us by tapping into shared human experiences and emotions. You can help stakeholders relate to your perspective and understand your motivations. Storytelling is a powerful tool in your arsenal to win a lot of battles. It is a gift to most who demonstrate it, however, with hard work, you can get much better.