AV

Arjuna Velayudam

Full-stack engineer · New York, NY

velayudamarjuna@gmail.com LinkedIn GitHub Portfolio

Summary

Full-stack engineer with a deals-transformation background. Build AI agent systems and data products end-to-end — from schema and pipelines through model evaluation, UI, and deployment. Recent independent products: Foreman (AI revenue intelligence for home-services contractors) and Cairn (AI career operating system).

Experience

  1. Senior Associate, Deals Transformation · PwC

    – · New York, NY

    Lead deals transformation across PE diligence, separations, and synergy programs.

    • 25+ engagements across industries on PE standalone, separation, and synergy diligence.
    • Promoted first in class on a one-year accelerated timeline.
    • Led teams of 3–7 across US automotive and industrials on multi-workstream programs.
    • Shipped a commercial scenario hub; generalized into a solution offering driving $5M+ in revenue.
    • Built custom ETL aggregations to improve management forecasting and market-penetration analysis.
    • Led C-suite report-outs translating data into clear narratives.

  2. Data Analyst, Disputes & Investigations · Alvarez & Marsal

    – · San Francisco, CA

    Forensic analytics and visualization for legal damages cases — lost revenue, IP infringement, contract breach.

    • Identified $600M+ in questionable offshore investments delegated to tax havens.
    • Cleaned and evaluated multi-million-row SQL datasets for exhibit-ready outputs.
    • Built repeatable SQL pipelines and forecast models with documented assumptions and audit-ready traceability.
    • Delivered exhibits and written analyses for expert reports across damages and contract-breach cases.
    • Took projects from raw data dumps to multi-dashboard products using Python and Tableau on six-month timelines.

  3. Analyst, Private Wealth Management & Operations · Goldman Sachs

    – · New York, NY

    Analytics, compliance remediation, and reporting for PWM operations across 12 offices.

    • Cut a manual month-long annual process to a one-day system update.
    • Designed a 3-month remediation workflow for new regulations across compliance and advisors.
    • Built custom Salesforce reports analyzing advisor wins and incentive-driven growth opportunities.
    • Developed VBA automations to track fees and compensation across 12 offices.
    • Ran cost-benefit analysis on underfunded accounts to compare revenue gain vs management-fee loss over time.

Selected Projects

  1. Foreman

    Live ↗Repo ↗

    AI revenue intelligence for home-services contractors — customer scoring, reactivation, response classification, multi-channel.

    • Tone profiler agent learns operator voice from sent Gmail.
    • Scores past customers 0–100 by recency, LTV, frequency, and equipment age.
    • Reactivation analyzer drafts personalized outreach in operator's voice.
    • 7-category response classifier routes booking intent to calendar slots.
    • Email + Twilio SMS, Google Calendar integration, equipment lifecycle tracking.
    • Tracks revenue attribution end-to-end — outreach to booked job.
    • Stack: Python, FastAPI, SQLAlchemy, Twilio, Gmail OAuth, Google Calendar, Sentry, Railway.

  2. Cairn

    Live ↗

    AI career operating system — career graph, opportunity prep, agent orchestration. Multi-user, BYOK from day one.

    • Builds a deep persona from career materials; maps next moves as a graph.
    • Targets three personas: The Lost, The Leaper, The Prepper.
    • Surfaces: Command Center, Career Map, Pipeline, Profile, Admin.
    • BYOK LLM architecture; per-user usage logging from day one.
    • Stack: TypeScript, Next.js, pnpm monorepo, Railway.

  3. Azimuth Trading

    Live ↗Repo ↗

    Market intelligence platform for used oil & gas equipment — mud motors as the wedge segment, with a bronze→silver→gold data pipeline.

    • Aggregates public data sources to surface supply concentration, demand signals, and arbitrage opportunities.
    • Medallion data architecture: LLMs extract stated facts with quotes; deterministic Python rules produce inferred facts.
    • Bipartite globe (supply ↔ demand) on the homepage with named-place anchors and basin polygons.
    • Provenance discipline: source, ingested_at, and confidence_level on every record.
    • Stack: Python, Postgres + PostGIS, Next.js, uv, Railway, globe.gl.

Education

  1. MIT xPRO

    – · Remote

    Python for AI & Data Bootcamp

    • End-to-end ML pipelines, computer vision, model evaluation.
    • Capstone: malaria-detection CNN with explainable evaluation artifacts.

  2. Northeastern University

    – · Boston, MA

    B.S. Industrial Engineering, Math + Economics minors. Magna Cum Laude, Honors College.

    • 3.9 GPA, Dean's List.
    • Three 6-month co-op rotations: NYC, San Francisco, Mexico.
    • Coursework: Operations Research, Probability & Statistics, Linear Algebra, Engineering Databases.

Certifications

  • AWS Certified Cloud Practitioner (CLF-C02) · Amazon Web Services · · Verify ↗
  • Google Project Management Certificate · Google (Coursera) · Verify ↗

Skills

Languages
TypeScript · Python · SQL · JavaScript
Frameworks & runtimes
Next.js · React · FastAPI · Flask · SQLAlchemy · Tailwind · Node.js
AI / data
LLM application development · AI agents · RAG · BYOK architectures · Prompt engineering · Data pipelines · Statistical analysis · Computer vision
Cloud & infra
AWS (CLF-C02) · Vercel · Railway · Sentry · PostgreSQL · Alembic migrations
Integrations
Gmail OAuth · Google Calendar · Twilio SMS · Salesforce
Domain
M&A diligence · Deals transformation · Forensic analytics · Synergy modeling · PE separations