To make your question short, and to see if I understood it correctly, we're talking about a player who made his character a certain one and roleplays it entirely different. You added that you think that it comes from inexperience, and that he created this character after you said "no" to some "freak-character"-ideas. You want to help him roleplay the character he created.
As I see it, this problem is made from two smaller ones. The first is that he doesn't see his character as interesting because the character "is normal and normal is boring". The second is that you wanna help him understand why the way he plays the character does not fit the story-world of your game.
Helping him understand that "normal is not boring
This is the more important problem, as it stands in the basis of the entire problem. If he'll see that normal characters can be interesting his "anti-persona" will perish and he'll roleplay a normal character and not a freak one. The main trick here is to show him that normal characters are not entirely normal, i.e. "no person is like the others". In order for that to work, we need to give the character depth.
The easiest way to give depth to a character is through internal conflicts. Having goals and all is nice, but without something that blocks oneself from achieving them it is far less interesting. First thing to do is to go over his character's background and see if he implemented there an internal conflict for his character. If so, show it to him and talk with him about it. If no, sit with him and help him to come with one. The internal conflict doesn't have to be extravagant, but it needs to be there. An example one might be that he loves Vincent's sister but secretly hates Vincent himself, or another like Loves the sister but thinks that he's not good enough for there. I'll take the second one as an example for this section.
The conflict gives us a few things, a few added benefits. It gives the character 2 conflicting goals: "Get the sister and prove that I'm worthy". Now, with those two we also get a kind of an achieving-plan: "If I'll show her that I'm worthy, by getting something amazing done, she'll want me and I'll be able to get her". More than that, the character gets the knowledge that each advancement in order to achieve one goal will drive the other one to the far end.
But the first conflict is even more interesting. The character here has the knowledge that he needs the brother in order to save his lover, but he just can't stand being near the brother. He'll drive the mission onward for two reasons but he'll have doubts about his lover- if he'll marry her he'll be stuck with this brother of hers.
To make long story short, simple conflicts can show the player that even normal characters are interesting and unique. When combined with goals they force the character to take certain steps along the roads, to commit certain actions along the way, that he won't want to do but will make him doubt himself and question himself and see that his problem are far more interesting than those of every freak that he'll encounter.
Another nice way to help him see the importance of conflict is through showing him and analyzing with him certain protagonists that are normal people, from the stories and movies and series (of any form)that he likes. He'll see quite quickly that the conflicts make them interesting.
But he may say that it is not enough. For that there are a few more literary tools that might help him see why normal people are interesting. The first one is having flaws (internal or external) and the second one is using "The Ghost".
Flawed characters are characters that just like normal people aren't perfect. Those flaws can be internal (self-doubts, for example, or a mild paranoia) or they can be external (they're look frightens ordinary people, for once, or a missing hand for the other). The idea is that the character has to deal with the flaw, and one day to find the strength to overcome it. The fight for the overcoming act makes the character far more interesting. A nice example of that can be seen in The Rain Man, where he learns at the end that he can count on strangers/"dumb" persons like he's brother. Another nice example can be seen in the story of The Ugly Duckling who although looking terrible learned to acknowledge himself and to accept the way he looks, to accept his difference.
"The Ghost" is an event from the past that just like a ghost haunts the character to this day. Again, trying to cope with it is what builds a deep character. One example for this can be seen in the movie Inception, where we literally have a ghost- Cob's wife. Another example for this can be seen in the movie Casablanca, where he has to deal with his broken relationship with Ilsa. This Ghost is far more interesting as the originator of the Ghost actually comes back to his life. In Frozen we see another kind of a Ghost- the act that one feels guilty about. Elsa actually killed her sister.
All of these techniques are there for one reason- to make regular people interesting, to give depth to the characters, to make them human beings with goals and drives and psychology.
Helping him see that his character doesn't fit the world
After he understands that he doesn't have to be a freak in order to be interesting, he will be far more understandable about playing a character that fits the world. Then, try to explain to him as calmly as you can what it is in the way he played his character that doesn’t fit the world.
Explain to him that the characters are in a world where being a freak is bad, where achieving one's goals is the ideal. Each and every one for himself, as the saying goes. Give him examples from the way he played his character and analyze with him, in a one-on-one conversation where his way of acting came from. Use the background he created to illustrate to him where your problem comes from.
Then ask him what problems he has with his character, and together try to find a solution. Maybe let him be just a little bit freakish. Maybe he needs to just create a different character. This is basically between you and him. After that show the updated character to the group and get their approval.
When combining those two, you'll get a player who his far more willing to both play the character while also seeing the problems with the way he played his character before.
Combining the two solutions
When combining the two solutions you get a better player, who understands for the future also how to create regular characters that are not freaks yet far more interesting than those freaks will ever be able to be. Furthermore, you get a player who is willing to play his character as written while still making the character fit into the world. Hope any of these helped you.
Firstly, you should talk to all your players about the issue outside of a session. You can get some personal impressions first with one-on-one conversations, but ultimately the entire group should sit down to discuss the problems. Make sure the discussion is democratic in nature, though. JohnP points this out in a related question: "The group setting can be dangerous, as it can turn into people digging up old grievances or ganging up on a particular person."
During the discussion you should ensure that everyone is on the same page about how they want the game to be played. The same page tool is a useful set of questions that will drive the discussion in the right direction. Just make sure everyone is fully involved in the discussion and are voicing their opinions clearly (no passive-aggressive bs). If your players are open to compromises a consensus should be reached.
Secondly, make sure your friend is not suffering from My Guy syndrome. If the description matches (and it sounds like it does), show her the link privately and let her think about it. It should help her be more aware of her role as a player and hopefully remedy some of the issues.
You also mentioned your player gets upset as a result of your decisions. This related question contains a lot of suggestions for how to handle players that take things personally, ranging from studying your own approach to kicking out the problematic player.
In the end, though, you are the GM. You are the writer of half of the story, the referee on all mechanics, and the leader in the quest for fun. The way you drive your campaign is part of your style and your players should respect it. They need to be aware that your goal is always to increase the enjoyment they collectively get out of your game. A certain level of trust and respect is absolutely necessary. If this is impossible to obtain in your group, then the group as it is cannot function. Kicking out players or stepping down as GM would be the next steps to try.
Because the GM is so special, though, new players often fail to grasp just how complex the GM role can be, and can at times see him as an enemy and spoilsport. A neat "trick" you can use to show your players what being a GM is really about (that also gives you some rest from the responsibility of being a GM) is to have another person in the group be the GM for a few sessions. It doesn't have to (and most often shouldn't) be the same campaign you're running. Instead, it can be a few sessions of an off-shoot campaign. It's fun for the players because they get the chance to try out new (and often times silly) character builds, and the new GM will discover what it feels to have all this responsibility. Once everyone has GM-ed a couple of sessions, you will all have an idea of who's best at it and will work towards keeping that person as GM in the future. There's a chance it might not be you, but in the end it should result in a better experience overall.
Keep in mind that not everyone is fit for GM-ing, or willing to try at all. Don't force players to GM, and if they decide to try, encourage them to design very short adventures (no more than three sessions). They can always expand on them later if they like it, or end them early if they don't. The player that questions your decisions often probably thinks they can do a better job, so they're likely to accept your offer to prove themselves.
Best Answer
To solve an interpersonal issue at the table, I would use a strategy that effective mediators use to solve disputes between conflicted parties (even though you're one of the parties).
This strategy helps solve co-worker conflicts, relationship disputes, and even tabletop issues!
First, identify exactly what it is that you find problematic
Is it how the player is acting? Is it the characters backstory? Is it both? If you find both problematic, you'll want to address them separately.
To me, having a tragic backstory is not problematic in and of itself, but using surviving rape as justification to now having no respect for personal or property boundaries, being affronting to friendly NPCs, and mutilating corpses to collect trophies is downright insensitive. After identifying what you find problematic ...
Frame the issue in a way that you and the player share the problem
Since you both have a shared problem, you are both stakeholders in the solution. This is a tactical way of framing an issue so that you don't push the entire solution onto your player, and you don't take ownership of the entire issue either. You want to identify that you desire to work together to rectify the problem.
An example of framing the player's actions as the issue is something like "The party tends to have goals X, Y, and Z. Your recent actions of stealing from the NPCs and breaking into another NPC's house haven't helped reach that goal. Let's come together to find in-character methods you can act like a sneaky thief that also further's the party towards their goals." After framing the issue...
Propose a solution
The player wants to sneak around, break and enter, and be ghoulish. Your two most apparent options are to give them the appropriate playground to act this way, or to realign the character's actions into a different outlet.
The path of least resistance is to give them encounters that allow them to act this way without it being a detriment, and allow in-game information to broadcast that they're the best character for the job. Tweaking a premade campaign to highlight a specific character type is encounter-specific and hard to generalize, so if you want to go down this avenue and need help, that'd be best for a new question. This approach allows the player's character to at how they want, identifies the GM's environment in which they're acting as an issue, and could give them positive feedback from the party for playing their role effectively. This method puts the solution squarely on the GM's shoulders. As another player, you should communicate your issue with the problem PC to the GM and see if they're willing to adapt the campaign in this manner. Alternatively, you can also communicate in-game when you're wanting the player to sneak around and gather something. Including them in the goal-making process and having them give feedback in the strategy helps them become included and offer their skillset in a productive fashion.
The path of more resistance is to reel the character back in to act more in line with the party's current expectations. You can do this directly by telling the player that what they're doing is problematic and antithetical to the campaign, or you can do this indirectly by having in-game consequences to their actions, such as guards showing up to detain the PC, NPCs visibly disturbed with the party for consorting with the problem PC, or, since you're running Curse of Strahd, having the town liken the PC with Strahd and his legion because of their actions. Either direct or indirectly identifying this, this method puts the solution squarely on the player's shoulders. As another player, you can privately reprimand the player's actions in-character or try to buddy them up with someone that can help monitor and assist the character to accomplish their personal goals without necessarily breaking in-game law.
In reality, the solution is likely somewhere between those two ends of the spectrum, and getting the player on board without feeling personally attacked is the way to move forward without hurt feelings. You know your player more than we do, and you know what communication may or may not be effective.