Pranjal Sailwal
recovering from log
The same work, ordered for the role you have open. Pick one: you get the projects to open first, two figures you can check yourself, a résumé made for that role and an email with the subject already written. Every figure carries a stamp that says how it is known.
I start with the question, write the decision down with its reason, and show how each figure is known. These are the projects that show it from the first question to the measured result.
98.1%
of the cost increase happened after the click
11 of 11
planted changes caught in a live pass, no false positives
Type a skill, or paste the job description. It runs in your browser and nothing is sent anywhere.
Drag to move from what the headline said to what the work found. Each figure keeps its stamp.
Spend went up, and the cost of each conversion went up with it. So spend must be the problem.
19.2x
more spend in the larger campaign
Splitting the cost into its three factors showed where the increase really happened: after the click, where conversion collapsed.
98.1%
of the cost increase happened after the click
73.8%
drop in post-click approved conversion
One smooth number for the neighbourhood: the readings averaged, the gap between them gone.
The gap itself, and a confidence that falls as the sources disagree. A missing value is shown as unavailable, never guessed.
10 to 20%
gap between sensors placed side by side
11
data sources, each with licence, rate limit and how provenance is kept
Technical product and programme work: technical product management, product analysis, associate product roles, programme management, and the consulting and implementation roles next to them. I want to be on a team where the question is asked before the build, and where writing things down is normal, because that is also what remote teams run on. I claim no title yet. The projects are the evidence.
Three habits, each with something you can check. I start from the number that does not add up and put a figure on the problem before I design anything. I write the decision down with its reason, so the next person does not have to guess. And I say how each figure is known: proven, documented, or only my word. This site is built the same way.
98.1%
of the cost increase happened after the click
15
documents in its docs folder, from the product brief to the limitations file
3
design documents: a blueprint before the code, a verification matrix and an interview guide after
Turning a vague problem into a question I can measure, then checking the answer against something known: an analytical result, a planted change, a significance test. I would rather show a result that survived a check than a bigger one that did not.
11 of 11
planted changes caught in a live pass, no false positives
68
unit and integration tests
The growth analytics lab. A campaign spent far more and converted far worse, and I wanted to know which lever actually moved the cost. I built a three-factor bridge that leaves no residual and found where the increase happened, using only public data. It started with a number that looked wrong, which is how most of my work starts.
19.2x
more spend in the larger campaign
98.1%
of the cost increase happened after the click
The part I enjoy most comes before the code: working out which problem is worth solving, who it is for, and how we will know it worked. I can build, which means I know what a decision costs, but I want to spend my time on the decisions.
I have not held a product title or shipped to paying users, and my grades are solid rather than top of the class. What I have is a set of projects built end to end, each with its limits written down. I would rather show the work than ask you to take it on trust.
7.36 / 10
CGPA over eight semesters, no backlogs
I write things down, so people can disagree with the reason and not only the result. On the air quality platform I worked in a four-person team, and its limitations file says plainly what the platform cannot do, which is what lets a team trust it.
4
people on the team, about 1,050 developer-hours, Pranjal as backend lead
Press Audit me on the home page: every figure opens its source, the command that reproduces it, when it was last checked and whether its seal still matches. The GitHub page shows what GitHub itself says about the work, and each certificate has its credential ID.
One page to forward, and the résumé in three formats.