the hardest thing for us (humans) to do at the moment, is to alter our own cells to do specific things; and our lungs are sterile environments. They do not tolerate infestations, and bacteria that do manage to survive and colonize our lungs are usually fatal (or sickening, like TB which was once fatal and is now only severely debilitating.)
So, if you are looking for a bacterial approach to solving the self-poisoning of humans by cigarette smoking; then you would need to research the ones that live in our mouth; and even then, you would not be successful, because the amount of surface area in the mouth versus the amount of smoke inhaled per puff from a cigarette, would only negate about 4% of the tar inhaled (nobody who smokes cigarettes sits there and swishes cigarette smoke around in their mouth for 60 seconds to make sure the negligible amount of bacteria living on their teeth has had a chance to metabolize the poisons in the smoke.)
The idea you have is great: these are the kinds of ideas that truly push synthetic biology forward. it is a creative approach for applying bacterial metabolism to a known human poison; cigarette tar. But you must realize the cause for this human poison is human activity-based: the amount of engineering involved in producing a custom bacteria that will somehow consume the tar build ups in human lungs and have zero pathological impact is incredibly large, and impractical when compared to stopping humans from smoking in the first place. e-cigs, and their oil vapor based nicotine delivery systems completely bypass the tar problem already. We have cigarette butt build ups in our gutters because we still have an amount of cigarette smokers due to humans with a built up habit that is not able to be overcome by a new technology (see: old dog and new tricks.)
If you want to engineer a bacterium that can live in our human lungs and consume/destroy incoming poisons/pathogens, then you are also engineering a symbiotic organism that should also have a cousin that lives on our skin and destroys any other foreign pathological bacteria. And when we do that; if the friendly, engineered bacteria is not somehow directly linked to living off our human skin, it will spread to live on other surfaces, and then, it will destroy 99% of the other bacterial species in existence; and the remaining 1% will have a supremely lethal effect on us, the humans, because we will have no immune system bases to work of (unless we simultaneously manage to continuously vaccinate ourselves.)
so perhaps, the question is not whether we can engineer a bacterium to eat the tar in our cigarette smoker's lungs; but whether it is worth it, in the long run for humans as a whole (and as an extension, the entire biome we inhabit.)
I really like the posited question of how we manage to modify our own constructs to deal with imbibed poisons (ie: how do we alter our own cells such that our lungs are able to disperse of toxins such as tar, before it builds up into a problem) I am all for human regeneration "perfection", but we do not reach that plateau by banging our heads against every wall along the way (or rather, wearing helmets the entire time. That means, it will be worthless to learn to overcome tar problems without learning to overcome the smoking habits first. I suggest in becoming familiar with "the helmet rule" to better ascertain an understanding of human behavior, and the inherent physiology.)