## GSoC Mid-Eval Review

### Mentoring Feedback

- Generally very responsive, mostly get feedback/resolution within a day (sometimes PR reviews take a bit of time which acts as a blocker, but usually is resolved after a reminder or two)
- Understanding and patiant, helped a lot in getting used to initial sktime environment (which makes me doing work a lot faster now)
- A professional workflow - not really a negative or a postiive, its just a tougher change since I am mostly not used to it (majorly worked with other uni friends etc :p)

### Project Goals

- Initial project goals were divided into 3 terms (based on the work required):
- Term 1: Creation of DL Forecastors and Integration of sktime-dl (major work component)
- Term 2: Adding other forecastors, and implementing DL Classifiers if required
- Term 3: Adding soft features and completing documentation
- My Progress:
- Although the tasks have been followed, schedule is not exactly going as projected
- Migration of sktime-dl is almost completed, it has become a good first issue now which can be easily picked up by new developers.
- Have been adding soft features throughout, which is close to completion as well (need to finalise design)
- Major work left is design and implementation of DL Forecastor, which will be my main goal in the second half of the GSoC.
- Personally, I am happy with the amount of work and changes I have contributed to, especially on a design end (which has made it easier for others to contribute and help as well).
- I feel I should comfortably able to finsih the design and implementation of basic DL Forecastors (although design will require help)

### Learning Goals

- Contributing to open source, and large code base, in a way which is understandable by other developers and easy for them to develop upon it in the future.
- One of the goals I am most satisfied with, learning how to use test environmnents, write documentation and especially the sktime-dl merge issue so that users are able to use it as reference and add code to sktime has been a good learning experience in total.
- Understanding best code practices in multi-developer environment
- To be honest I felt comfortable enough writing good code for multi-developer environment, but my main learning experience was wrt communication in community. They way I present instructions, or ask doubts related to code, or present my design had been initially with me in mind, with the idea that I can explain it to someone if they ask, but past weeks have helped me learn how to make myself clear, and convey all information required in a better manner.
- Improving upon my knowledge of sequential Deep Learning models by exploring different approaches and understanding the mathematical insights behind them.
- This hasnt been dealt with much, I learnt more about design than particularities of DL models (although here and there I read the research paper specific models were referring)
- Hope to learn more about this in the second half