When going through your game with the help of an engine, I think it's absolutely key to keep a human perspective and try to understand why a certain move is better. For instance here the reply 2.Be2 is not human at all, you must really check why 2.Bf3 and especially why 2.Bf7 doesn't work. Maybe you did this but it is a very common mistake, and should be kept in mind at all times when working with an engine.
Allowing the Maroczy is not a good idea IMO, after 3.Nf3 I think the best move is 3.-e5 followed with a Botvinnik-style setup with g6, Bg7, Nge7, d6 and f5. If White goes g3 his N will be misplaced on f3 as it blocks the f-pawn (he should develop the B to e2). 3.Ne2 is a better move though, maybe Black's best then is to go for 3.-Nf6 4.Nbc3 e6 5.d4 (otherwise B gets d5 in) cxd4 6.Nxd4 Bb4 with a Taimanov line that should be fine for him.

Of course I didn't, otherwise I would have written the wonderful lines I'd certainly have a hard time spotting them in blitz:

Thanks for the opening advice too!

Yesterday, I faced another unconventional Sicilian challenge vs the chess.com comp - a misplayed Smith-Morra that I didn't make use of (i.e. didn't extract an advantage out of the opening) I guess that, after 1. e4 c5 2. d4 cxd4 3. Qxd4? Nc6 4. Qd1 Nf6 (all std so far) 5. Bc4, I misplaced my own bishop onto c5 - I thought it would be useful in a kingside attack; but the 'theoretical' move is Bb4: I guess, its purpose is to prepare d7-d5 to get rid of this weak pawn (unlike Guioco Piano, there's no c-pawn that would support it on d6, plus the d-file is semi-open), and also White can exchange this bishop on e3 (which did happen).

I eventually sacrificed both the d6 and e5 pawns to let my rook come onto d2 (because I knew that being down on time as usual with equal material would result in a loss vs the comp and wanted to go all-in), but the problem was that White's Rad1 woulda shut the activity down; fortunately, the comp played 20. Rac1?? instead, leading to the following position:



Here, Qf2 is an obvious mate (on either h2 or f3), how can I be so blind?

We continued with 20... f5?? 21. Qc7?? g6?? (this was downright stupid, g7 was covered by the knight anyway and there was no need to shut the g-rook down so far; Qf2 still works here) 22. Qxb7?? (after Qe5+, White is just two pawns up) and I finally implemented a spectacular faster mating idea that I had seen a couple of moves before but couldn't implement then: