diff --git a/content/_index.md b/content/_index.md
index aa60114..c61cc71 100644
--- a/content/_index.md
+++ b/content/_index.md
@@ -4,21 +4,22 @@ title: home
 <img
   id="personal"
   src="/images/me.jpg"
-  alt="My face">
+  alt="my face"
+  title="i am aware that my face in not aligned with the curve of the arch behind me. your complaint has been duly noted and promptly ignored.">
 
-# Vasilis Valatsos<br/>(Βασίλης Βαλατσός)
- 
-software engineer @ [new odyssey](https://new-odyssey.com)
+# Vasilis Valatsos<br/>(Βασίλης Βαλατσός) 
+
+## software engineer @ [new odyssey](https://new-odyssey.com)
 
 experience as a backend developer and data scientist
-- <span style="font-family: monospace;">—</span> implemented postgresql databases and relevant apis.
-- <span style="font-family: monospace;">—</span> built data pipelines 
-- <span style="font-family: monospace;">—</span> provided insights for decision-makers
-- <span style="font-family: monospace;">—</span> [influenced policy making](https://contrapodernews.com/jorge-jraissati-denuncia-a-reguladores-de-criptomonedas-tuvimos-que-confrontarnos-a-la-instrumentalizacion-politica-de-las-regulaciones/) in the EU with data.
-- <span style="font-family: monospace;">—</span> helped a non-profit organization [win prestigious awards](https://www.atlasnetwork.org/stories/african-solutions-for-african-problems)
+- implemented postgresql databases and relevant apis.
+- built data pipelines 
+- provided insights for decision-makers
+- [influenced policy making](https://contrapodernews.com/jorge-jraissati-denuncia-a-reguladores-de-criptomonedas-tuvimos-que-confrontarnos-a-la-instrumentalizacion-politica-de-las-regulaciones/) in the EU with data.
+- helped a non-profit organization [win prestigious awards](https://www.atlasnetwork.org/stories/african-solutions-for-african-problems)
 
-now focused on bringing reinforcement learning back in the mainstream, combining it with state-of-the-art natural language processing methods, such as large language models.
+focused on bringing reinforcement learning back in the mainstream, combining it with state-of-the-art natural language processing methods, such as large language models.
 
-m.sc. in mathematics (statistics and data science),<br/>
+m.sc. in statistics and data science,<br/>
 university of turin,<br/>
-[m-thesis](/pdf/mthesis.pdf): implemented ppo in a custom environment named pneuma
+[m-thesis](/projects/master-thesis): implemented ppo in a custom environment named pneuma