Balancing Variety

Alrighty, here goes my personal opinion:

First of all, planning to win by flatline is a bad idea, in general - it's a decent plan B, and the threat of flatlining can be a very effective pressure tool on the Runner, but actually pulling it off requires either pretty big misplays on the Runner's part, or a very lucky coincidence.

What Jinteki is very good at is the pressure part of the flatlining threat, the actual kill though? Not so much. That's why you should always be planning to ultimately win via agendas.

Now, the questions:

Ices, any interesting alternatives?

If you're going for the variety aspect, I feel like you're missing Snowflake. Regardless of how it appears at first glance, it's basically a super-cheap Wall of Static if you play it right - only unlike most other barriers, it's at its strongest once the Runner actually has the right icebreaker. Then, the trick is to play Snowflake so that they won't want to risk ending the run and will pay more to actually break it. Then, it starts being awesome - if they have Corroder, you paid 1 to rez (and maybe 1-2 to install, if it's layered) and they're paying 2 to get through every time.

I'd drop the second Ice Wall and upgrade Heimdall to 2.0 - that version is much more reliable re: actually landing the brain damage. Anyone worth his own weight in water isn't going to run Jinteki on their 4th click... but they just might on the third. If that's the case, Heimdall 2.0 will stick 'em with 2 BD, which pretty much makes Fetal AIs (and by extension, any agenda installed and advanced twice) super-easy to score.

In the code gate department, I'd drop the Sensei for a third Chum. It's an enourmous pain for people that are packing Yog and no Ice Carver, makes other ice a must-break (Ichi), can be used to set a cheap kill (depending on if they're rocking a dedicated suite or mainly AI stuff, either Chum a Swordsman or a Whirlpool and things become interesting), and is great early-game stopping ICE - most people will jack out after encountering a Chum unless they have a full breaker suite.

Sensei, on the other hand, is pretty "meh" - it costs more, and doesn't actually do all that much - only makes getting through the ICE after it more expensive by 1 credit. If you install Sensei in a shallow server (maybe 1 or 2 ICE after it), most people will just shrug and take that penalty, as it's cheaper than breaking through Sensei. But that's the thing - Sensei only becomes good with a deep server, then it costs a fortune to install a rez (we're talking 6-7 credits, easily - since you're not packing Interns) and every runner has to have a plan for STR 5 Code Gates (because Tollbooth is a thing), so you can't even plan on it actually working its magic at that point. In short: don't play Sensei :P

If you want to go really funky, you could find place for one Cell Portal. Don't expect to use it with any kind of regularity, just have your eyes open for all the cool plays you can make with it if the stars align. And, if they don't, you can use it to trash some Caissas :P

You're pretty much covered on interesting sentries, aybe with the exception of Shadow, although I don't know what I'd drop for it. As far as Traps are concerned, you really want a Data Mine - again, it's super-dangerous with Chum (my favorite server is Chum - Data Mine - Snowflake, actually), cheap to rez, and overall just cool.

Economy, is there enough to maintain solid pool of credits?

Try it out and see. It might, or it might not. My guess is you'll be mostly fine, although I would add the third Melange (if you have 2 core sets - the list looks like maybe you're working with one only). If you find you need more economy, the card to add here would be 3x Interns. It's super-awesome for recycling Melanges into a taxing server (because you can do it even when almost broke, as it only costs you actions and not credits), and can also be used to recur Hokusais, Troubleshooter, trashed Jacksons and even an agenda (or ambush!) that Noise milled and didn't see yet. Hell, you can re-install a Whirlpool or Data Mine with it, for bonus points :P

What would be the 5 cards you would take away to make it 49 card deck?

This goes against conventional wisdom, but my answer here is going to be a bit unconventional: "I wouldn't, I'd add cards instead".

I would actually change one of the Niseis into a Priority Requisition and add another 5 cards. Do the math on Agenda and Ice density depending on what cards you add, but I wouldn't be surprised if you find it actually going up when you do that. It's something I'm trying in my current Jinteki deck, and so far it's been very solid - much more so than conventional wisdom would lead you to believe. Especially now that Keyhole is all the rage, and Noise has started rearing his ugly head again.

(sidenote: one of those cards I'd add would be a third Neural EMP)

What would be alternative influence options?

We pretty much covered that in the ICE suggestions, I think. If you had 5 influence to work with, I'd say add 3 Shadows and 2 Closed Accounts - it's recurring Siphon protection and is enabled by Snares. If you had another 2 influence on top of that, I'd add one Data Raven.

Those would be the answers off the top of my head, I'm sure there's plenty other viable choices/answers though.

These are really great inputs. Thank you. To sum it up:

Ices: -1 Heimdall 1.0, +1 Heimdall 2.0 -2 Ice Wall, +2 Snowflakes -1 Sensei, +1 Chum

Potentially,

+1/2 Mine Traps +1/2 Shadows +1 Data Raven

Agendas

-1 Nisei, +1 Priority Req

Others

+1 Neural Emp, +1 Melange, +3 Interns, +2 Closed Accounts

I love Shadows but with so many sentries I felt compelled to add barriers and Ice Walls were tempting for either ToL adv. tokens storage, eating Crypsis virus counters, fending off early Atman with no Datasucker, or face checks on Asset servers. I do agree Snowflakes are more taxing in the long run, though. If I drop Sensei on top I'm left with not a single solid ETR, ouch.

Even though it goes against principle of this deck adding 3rd Chum makes sense. What stands behind Chum is relevant and if it is most likely something deadly, Chum will work as ETR.

I had Cell Portal in the early version of the deck. It feels to me as too hard to pull off, and probably more belonging to Replicating Perfection with +1/2 copies of Sunset.

So in the end, the additional cards would be: +1 Shadow, +2 Data Mine, + 1 Closed Accounts, +1 Interns.

Thanks again.

I love Shadows but with so many sentries I felt compelled to add barriers

Here's the funny thing - sentries actually become more powerful when you have a disproportionate amount of them. Most, if not all have an effect that can't be (easily) ignored, and they're a bitch to break. If you stack sentries hard enough, Criminals running Faerie as their only dedicated Sentry breaker (with Crypsis to fall back on) become extremely sad. Similarly Mimic with Datasucker support - if you run a ton of sentries, it's easy to create a situation where getting sucker tokens is always a costly proposition.

And don't even get me started on what a 3-sentry server does to Cloak & Dagger setups :P

If I drop Sensei on top I'm left with not a single solid ETR, ouch.

That's one of the many ways Jinteki is weird. You don't actually need ETR ice to stop the runner. It takes a while getting used to, but actually works rather well.

Just think about it this way: ETR ice only ends the run if they can't break it (so basically just when you rez it), and then they often just bounce off of it, with no real harm done. So really, it's only bad for them if they desperately need to get to whatever's behind it.

non-ETR ice, on the other hand: usually does something nasty right away, and it's usually nasty enough that they won't run through it until they have a proper breaker. So, what you're really doing, is trading one instance of an ETR effect for something nasty. Often, the effect is nasty enough that letting them into the server is actually a good trade.

Add to this the fact that you're Jinteki (and the Core variant, no less), where every card in a server has the potential to do 1-3 damage, and things suddenly start to become extremely tricky for the runner. If you run face-first into a katana and are left with 2 cards in hand, are you really going to access? If it's a Fetal AI (or worse, a Snare or a Junebug), that could be game right there and then :) Add to this the (real or perceived) threat of Neural EMP (and Punitive Counterstrike, as of True Colors) and suddenly the danger zone (zing!) increases by a LOT.