Zero to One

Selected Work

Verrik
01

Verrik

2026

The independent trust layer for AI decisions in regulated industries.

DealLens
02

DealLens

2026

Multi-agent due diligence — six specialist agents, one risk-scored memo.

F1 RaceSim
03

F1 RaceSim

2025

Real-time race strategy simulation with AI insights.

Ritematch
04

Ritematch

2024

AI job matching and résumé tailoring with NLP match scoring.

All Work

I take AI products from a blank repo to deployed.

Work

Verrik — Independent trust layer for AI decisions

Verrik — Independent trust layer for AI decisions

2026
Problem
In regulated industries, carriers are liable for AI-agent decisions — even vendor models — yet have no independent way to prove a decision was reproducible, fair, and audit-ready before the EU AI Act high-risk deadline.
Built
A "flight recorder and auditor" for AI agents: captures every decision as a durable, tamper-evident trace (SHA-256 hash chain) with deterministic replay, runs four-fifths disparate-impact fairness evaluation in shadow mode, and auto-generates regulator-ready audit reports. Shipped as a TypeScript monorepo with a Postgres-backed Fastify API and a Next.js site.
Outcome
Reproducible, regulator-ready audit trails for autonomous AI decisions — independent of the agent vendor being audited.
  • TypeScript
  • Next.js
  • Fastify
  • PostgreSQL
  • Turborepo
  • AI Governance
GitHub →Live Demo →
DealLens — Multi-agent AI due diligence

DealLens — Multi-agent AI due diligence

2026
Problem
Due diligence means correlating evidence across finance, legal, tech, GTM, ops, and people under heavy time pressure — slow, manual, and easy to miss cross-cutting risk.
Built
A multi-agent platform where a planner agent fans out to six department specialists and a synthesizer cross-correlates evidence-cited findings into a risk-scored investment memo — with cost-aware model routing across providers and real-time SSE streaming.
Outcome
Evidence-cited, risk-scored investment memos generated end-to-end; deployed full-stack on Vercel and Render.
  • Python
  • FastAPI
  • LangGraph
  • Next.js
  • Claude
  • Multi-Agent
GitHub →Live Demo →
Phishing Shield — Agentic email threat detection

Phishing Shield — Agentic email threat detection

2025-2026
Problem
Phishing detection in Gmail/Outlook needs to be real-time, accurate across attack categories, and resilient when offline.
Built
Agentic Chrome extension + FastAPI backend using Claude 3.5 Sonnet for live classification (Phishing / Tech Support Scam / Scareware / Benign). MutationObserver scans incoming mail; local heuristic fallback covers offline mode.
Outcome
Sub-second threat analysis across 50+ test emails with confidence scores and color-coded toasts.
  • Chrome Extension
  • FastAPI
  • Claude 3.5 Sonnet
  • Email Security
  • Real-time
GitHub →
F1 RaceSim — Real-time race strategy with AI

F1 RaceSim — Real-time race strategy with AI

2025
Problem
F1 teams reason about pit windows, tire compounds, and weather under heavy time pressure. Existing public sims lack live AI commentary and multi-strategy comparison.
Built
Full-stack simulator (Next.js + FastAPI + Gemini) with multi-strategy comparison, weather-aware predictions, and interactive race visualizations. Deployed on AWS Lambda + Vercel with serverless optimization and AI inference rate-limiting.
Outcome
40% faster cold-start time post-optimization; real-time AI insights served at race-grade latency.
  • Next.js
  • FastAPI
  • Gemini
  • AWS Lambda
  • Vercel
  • Python
GitHub →Live Demo →
Ritematch — AI job matching + résumé tailoring

Ritematch — AI job matching + résumé tailoring

2023
Problem
Candidates send the same résumé to dozens of roles; ATS systems reject what should be relevant matches due to keyword mismatch.
Built
React + Django portal with REST APIs for NLP-based job-résumé similarity scoring and GPT-4o-driven résumé rewriting tuned per JD.
Outcome
20% improvement in candidate-job match relevance vs. baseline keyword matching.
  • Django
  • React
  • GPT-4o
  • REST
  • LLMs
  • NLP
GitHub →Live Demo →
Verbizz — Real-time business recommendations

Verbizz — Real-time business recommendations

2024
Problem
Recommendation latency kills conversion; batch pipelines can't react to live user signals.
Built
Cloud-native microservices platform on gRPC + Kafka + Redis + Flask with content-aware recommendation engine. PostgreSQL for warm storage, Kubernetes for orchestration.
Outcome
Scalable AI-first architecture serving real-time personalized recommendations.
  • Flask
  • Kafka
  • Redis
  • gRPC
  • PostgreSQL
  • Kubernetes
  • Docker
GitHub →Live Demo →
PitStopAnalytics — F1 Championship Predictor

PitStopAnalytics — F1 Championship Predictor

2024
Problem
Predicting F1 championship outcomes requires modeling driver, constructor, and circuit interactions across noisy historical data.
Built
ML pipeline (clustering + PCA + classification) over Ergast F1 API data exploring driver and constructor patterns.
Outcome
Interpretable championship-outcome models with visualized driver/constructor archetypes.
  • Python
  • Pandas
  • Scikit-learn
  • PCA
  • Clustering
  • Jupyter
GitHub →Live Demo →
BizStream — Real-time recommendation engine

BizStream — Real-time recommendation engine

2025
Problem
Recommendation APIs needed sub-200ms latency at scale; sync data pipelines bottlenecked I/O.
Built
Real-time recommendation system using Kafka + Redis + FastAPI with Docker Compose orchestration and async data pipelines.
Outcome
35% API latency reduction through smart caching and efficient async I/O.
  • Kafka
  • Redis
  • FastAPI
  • Docker Compose
  • Python
  • Async
GitHub →
LLMForge — Production sentiment analysis API

LLMForge — Production sentiment analysis API

2025
Problem
Need a deployable sentiment-analysis service with monitoring, framework portability, and stable accuracy.
Built
FastAPI + BERT service with Prometheus monitoring, Dockerized deployment, and PyTorch/TensorFlow portability.
Outcome
98% accuracy on IMDb; production-ready monitoring and dual-framework support.
  • BERT
  • PyTorch
  • TensorFlow
  • FastAPI
  • Docker
  • Prometheus
  • NLP
GitHub →
Formula 1 Insights Dashboard

Formula 1 Insights Dashboard

2025
Problem
F1 race analytics scattered across formats; no unified interactive view of driver, team, and race-level data.
Built
Streamlit dashboard with embedded Tableau views for interactive exploration of F1 statistics.
Outcome
Unified real-time view of race data, driver performance, and team analytics.
  • Streamlit
  • Tableau
  • Python
  • Data Visualization
GitHub →Live Demo →

About

I'm an AI engineer and 0→1 product builder — currently Software Engineer II at Adobe, building in-app AI experiences for Photoshop. Before that I was a founding engineer at early-stage startups, shipping LLM systems, multi-agent workflows, and evaluation frameworks into production.

What I love most is taking an idea from a blank repo to a deployed product. I built Verrik — an independent trust layer that captures, grounds, and audits AI-agent decisions for regulated industries — and DealLens, a multi-agent due-diligence platform that turns scattered evidence into a risk-scored investment memo.

Stack I reach for most: Python, TypeScript, LangGraph, Next.js, FastAPI, Postgres, AWS, and the Claude / OpenAI / Gemini APIs. I care about measurable outcomes — latency down, eval scores up, products in users' hands — and about the unglamorous engineering (observability, evals, fallbacks) that makes AI dependable in the real world.

Kiran Devihosur

Experience

Software Engineer II

Mar 2026 – Present

Adobe — San Jose, CA

  • Building in-app AI experiences for Adobe Photoshop — helping users learn features, discover tools, and complete creative workflows directly in-product.
  • Built an AI agent that automates unit-test generation for Discover Panel workspaces (+32% coverage) and a CI/CD coverage dashboard for test health and quality trends.
  • Designed a context pipeline and RAG workflow for in-product assistance, integrating Adobe Firefly MCP to assemble product knowledge, tool context, and workflow guidance per request.

Founding Engineer

July 2025 – Mar 2026

Anything (anything.com) — San Francisco, CA

  • Built model evaluation and debugging frameworks using AST-based semantic validation and complexity analysis across 400+ workflows, achieving a 78% autonomous resolution rate for adversarial agent outputs.
  • Engineered dynamic model routing and optimization logic across GPT-5, Claude, and Gemini, reducing inference costs by 10% while maintaining p95 latency under 800ms.
  • Reduced AI hallucinations and malicious outputs by 40% by instrumenting ML services with structured logging and metrics-driven observability to analyze adversarial engagement.
  • Architected multi-agent system contracts using Python and LangChain, enabling real-time, AI-driven security workflows for production customer environments.

Software Engineer

Aug 2021 – June 2023

Prime Therapeutics (via TEKsystems Global Services) — Bangalore, India

  • Developed event-driven, low-latency microservices using Java (Spring Boot), Kafka, and AWS to support high-availability decisioning systems handling 10K+ requests/min.
  • Implemented and evaluated XGBoost-based ML models for large-scale decision prediction, performing feature analysis and model validation to achieve 87% accuracy.
  • Conducted performance analysis and debugging across distributed services to improve throughput by 20% while maintaining strict correctness guarantees.

Associate Software Engineer

Nov 2020 – Aug 2021

Accenture — Bangalore, India

  • Developed user interfaces with React and TypeScript, increasing cross-browser compatibility by 25% and enhancing application usability.
  • Designed RESTful APIs in Python to streamline data processing, improving system communication and reducing processing time by 30%.

Education

Master of Science in Data ScienceGPA: 3.97 / 4.00

Aug 2023 – May 2025

University of Colorado Boulder — Boulder, CO

  • Data Structures & Algorithms
  • Machine Learning
  • Data Mining
  • Data Center Scale Computing
  • Cybersecurity for Data Science
  • Data Science as a Field
  • Ethical Issues in Data Science

Bachelor of Technology in Computer Science and EngineeringGPA: 8.52 / 10.00

Aug 2016 – June 2020

KLE Technological University — Hubli, India

  • Data Structures and Algorithms
  • Software Engineering
  • Web Technologies
  • Cloud Computing
  • Database Systems
  • Operating Systems
  • Object-Oriented Programming
  • Computer Networks
  • Distributed Systems

Tech Stack

Languages

  • Python
  • Go
  • TypeScript
  • JavaScript
  • SQL
  • HTML
  • CSS
  • Scala
  • R

Frameworks & APIs

  • React
  • Next.js
  • Node.js
  • Tailwind
  • FastAPI
  • Flask
  • Django
  • Spring Boot
  • JUnit
  • REST
  • gRPC
  • Kafka
  • GraphQL

AI/ML

  • GPT-5
  • GPT-4
  • GPT-4o
  • Claude Sonnet-4
  • Claude Sonnet
  • Gemini
  • LangChain
  • LangSmith
  • LangGraph
  • LlamaIndex
  • MCP
  • Pinecone
  • XGBoost
  • PyTorch
  • HuggingFace
  • TensorFlow
  • Transformers
  • scikit-learn
  • BERT
  • Reinforcement Learning

Cloud/DevOps

  • AWS
  • AWS Lambda
  • AWS S3
  • GCP
  • Azure
  • Docker
  • Kubernetes
  • Git
  • GitHub
  • GitLab
  • GitHub Actions
  • CI/CD

Monitoring

  • AWS CloudWatch
  • Prometheus
  • Logging
  • Alerting

Databases

  • PostgreSQL
  • MySQL
  • MongoDB
  • Redis
  • DynamoDB
  • BigQuery
  • Pinecone
  • GraphQL

Certifications & Awards

Generative AI FundamentalsDatabricksApr 2025Data Science HackathonUniversity of Colorado BoulderDec 2024Introduction to R ProgrammingUniversity of Colorado BoulderJul 2023MongoDB – The Complete Developer's GuideUdemyApr 2023Spot Award – Outstanding Contribution in ProjectTEKsystemsDec 2022Go Java Full Stack with Spring Boot and ReactUdemyApr 2022React Testing library with JestUdemyApr 2022React CourseUdemyMarch 2022SQL (Basic)HackerrankJun 2021SQL (Intermediate)HackerrankJun 2021Vue CourseUdemyApr 2021Python (Basic)HackerrankSep 2020Python (Intermediate)HackerrankSep 2020Python CourseCoursera & GoogleAug 2020

Volunteering & Activities

Officer

Aug 2024 - May 2025

CU Hiking Club — Boulder, Colorado

Serve as a Hiking Officer for the CU Hiking Club, overseeing the organization and execution of group hiking activities. Responsibilities include leading hikes, planning routes, ensuring safety protocols, and coordinating with club members to foster teamwork and outdoor exploration. Actively contribute to building a vibrant hiking community within the university by encouraging participation and promoting environmental stewardship

Volunteering

Apr 21, 2024

University of Colorado Boulder — Boulder, Colorado

I have given an impact of 6 hours of volunteering service for the Volunteer Resource Center during National Volunteer Week 2024 at the University of Colorado Boulder. My efforts were recognized and verified by the university as a meaningful contribution to the community.

Contact

I'm always open to discussing new opportunities, interesting projects, or just having a chat about technology and development.