EXPERIENCE / EDUCATION / PROJECTS

The work behind
the work.

RÉSUMÉ

Krishna Koushik Kasarla.

Hyderabad, India

01 Education

Computer Science & Engineering

Bachelor of Technology

Gokaraju Rangaraju Institute of Engineering and Technology

Hyderabad, India

Expected May 20298.44 / 10 CGPA

02 Experience

Freelance Software Developer

Self-employed via Fiverr ↗ · Remote, project-based

2020 – Present
  • Completed 26 paid client orders across Linux VPS and cloud setup, application deployment, and website development with WordPress and modern web frameworks.
  • Configured, deployed, and troubleshot client environments, including custom Minecraft servers, earning a 4.9/5 rating across 21 public reviews.

03 Selected projects

visionOS

AI-native spatial workspace

2026

Next.js · TypeScript · Gemini · tldraw · Web Workers · Kiwi.js · IndexedDB

  • Built a local-first AI canvas that turns sketches and natural-language requests into interactive apps, documents, and presentations.
  • Developed a Web Worker layout engine using R-tree indexing and Cassowary constraints, solving 100-node alignment problems in 0.55 ms median across 100 benchmark runs.
  • Implemented sandboxed app execution with user-approved network permissions and IndexedDB workspace recovery, backed by 454 passing automated tests.

NovaSecurity

AI vulnerability analysis engine

2026

Python · FastAPI · Tree-sitter · NetworkX · LangGraph · Amazon Nova · AWS Bedrock

  • Built a Python static-analysis engine using Tree-sitter and NetworkX to trace potential source-to-sink vulnerabilities across functions, covering 29 untrusted input patterns and 19 execution sinks.
  • Developed an 8-stage LangGraph workflow for vulnerability scanning and AI red/blue-team review, with isolated Git workspaces for generating reviewable patch proposals.

HealthGrid

Synthetic clinical data platform

2026

Python · FastAPI · scikit-learn · NumPy · React · TypeScript

  • Built a CPU-only FastAPI backend using causal models to generate synthetic patient cohorts and longitudinal health records, producing 5,000 patients and 61K+ visits in a local benchmark.
  • Evaluated statistical fidelity, ML utility, and privacy risk against holdout data and a Gaussian-copula baseline, achieving 0.581 C2ST AUC and a +0.029 TSTR–TRTR AUROC gap on the demo dataset.

04 Technical skills

Languages
Python, TypeScript, JavaScript, SQL
Backend & web
FastAPI, Node.js, Next.js, React, REST APIs
AI / ML
PyTorch, scikit-learn, NumPy, LangGraph, LLM APIs
Cloud & tools
GCP, AWS Bedrock, Docker, Linux, Git, Tree-sitter, PostgreSQL, Firebase, Supabase

HACKATHON RECOGNITION

2nd place in healthcare.

Software Hackathon Convergence 2K26 · VNRVJIET, Hyderabad

Built HealthGrid with a four-member team during a 24-hour hackathon.