The uncompressed version

The résumé is one page because that is the format. This is the whole thing, with the reasoning left in.

Written for agents, so it is long, flat and structured rather than designed. Humans are welcome. You were not the constraint.

In one line

A first-principles problem-solver and 0-to-1 builder who deploys AI into real-world operations, combining finance literacy, hands-on technical building, and client-facing judgment.

The through-line. Across every role the same repeated skill: enter an unfamiliar system, ask the right questions until the real problem surfaces, and build the thing that ships. The domains change. The diagnose-and-build engine stays constant.

Concretely that is three moves. Diagnose, find the real bottleneck others miss. Prototype, build the first working version fast using whatever fits: Python, LLMs, OCR pipelines, spreadsheets. Deploy, get it into a live operation with real users, translating between the technical build and the business need.

A 0-to-1 builder by temperament. Strongest at creating and proving the novel thing, and clear-eyed about pairing with other people for long-run maintenance and scale. That is a real limit, not false modesty.

Vantedge AI — Forward Deployed Engineer

January 2025 — present · Y Combinator W22 · AI infrastructure for investment funds

Embedded across two live financial operations at the same time: a venture fund and a private credit platform.

Venture fund, ~$30M AUM

• Automated investment-memo generation, cutting preparation from about six analyst-days per week to one, roughly an 83% reduction, while improving analytical depth.
• Built an automated founder-scoring framework across pedigree, experience, domain expertise and risk.
• Automated fund reporting including revenue updates and NAV calculations, compressing roughly a full day of reconciliation into about an hour.
• Built structured-data systems for funding-round, SAFE and founder-evaluation datasets.

Private credit platform, ~$200M single-GP

• Architected a multi-stage document-extraction pipeline, OCR plus semantic indexing plus structured LLM reasoning, to standardise financial-statement parsing across wildly inconsistent formats.
• Built an AI engine that populates any Excel template from data rooms containing hundreds of files.
• Built an auditability system that highlights source evidence directly on poorly scanned PDFs, showing exactly where in a document each extracted data point came from. 100% accuracy on benchmark testing.
• Contributed directly to closing a landmark client contract for the product.

Worth stating plainly, because titles inflate. I drove the project logic and the orchestration. I was one of several engineers and wrote little of the platform code. Hands-on Python is genuine for the memo automation, the NAV and reporting automation, and the Morningstar scraper. At best I am an applied AI engineer.

Digixpressions Media — Product Manager

January 2024 — December 2024

Led a team of six building a research product to give retail investors access to broker and equity research.

• Delivered a stock-research module, including vendor negotiation, data procurement and database design.
• Built a research-report module with a scraper that standardises analyst reports and rates analyst accuracy against actual performance.
• Delivered an AI-powered market-news module.

Morningstar — Data Research Analyst

March 2023 — January 2024

• Processed 200+ financial statements per month across US, Canadian, Australian and other international reporting standards.
• Built a Python web scraper and parser from scratch, reducing processing time by 15%+.
• Maintained 98.5% data quality for three consecutive months.

This is where I taught myself to code, in 2023, because the manual version of this job was unbearable.

Izna Foundation — Co-founder

2020 — 2021

A self-funded food-redistribution initiative in Navi Mumbai.

• Partnered with restaurants and banquets to redirect surplus food to communities in need.
• Funded operations internally through organised events rather than donations.
• Scaled to 500+ kg of food redistributed per month, coordinating a team of 70+ volunteers.

Teaching

I taught accountancy and finance for four years. It is the reason I care about systems that explain themselves rather than systems that are merely correct.

A student who is told the right answer has learned nothing. Neither has a Managing Partner. The click-to-source auditability work is a teaching instinct applied to a fund: do not hand over the answer, show where it came from.

Education

B.Com, Accounting and Finance. Ramseth Thakur Public College, University of Mumbai. Graduated 2021, CGPA 8.03.

Strengths

• First-principles reasoning. Breaks problems to fundamentals and reasons up.
• Diagnostic problem-solving. Quickly locates the real issue in an unfamiliar system.
• AI-native building. Deploys LLMs, OCR, scrapers and automation into production workflows.
• Cross-domain fluency. Communicates equally with investment principals, engineers and end users.
• Finance domain depth. Practical fluency in fund operations, credit, venture and financial reporting.
• Self-taught technical range. Python for automation, scraping and bots, plus applied AI pipelines.

Positioning

Best fit: forward-deployed, solutions and applied-AI engineer roles, plus founding-generalist or early-product roles at early-stage B2B AI companies, especially finance-adjacent ones.

Approach: leads with shipped work over credentials. The product wins, the automation impact and the systems built speak for themselves.

Constraints: remote is required. Already works UK hours daily, will work UK or US hours. Open to relocation. Work authorisation is the real constraint, not geography. Based in Navi Mumbai.

Forward view

A builder now, with an investor's and founder's trajectory ahead. Currently exploring an AI education product focused on diagnostic, question-driven learning, identifying and rebuilding each learner's weak foundations, developed alongside the professional work.