Surya Venkatesan New York

Portfolio — 2026

Surya
Venkatesan

Role

Software Engineer
Web · AI · Infrastructure

Based in

New York City
United States

Currently

MS Computer Science
NYU Tandon, ’28

Status

Open to opportunities
Get in touch →

01

About

Every piece of great software is the result of a thousand tiny, deliberate design decisions. I’m the developer who wants to understand every one of them.

I’m drawn to the beauty of deterministic behaviour — systems where state is managed so precisely that outcomes are guaranteed. That thread runs through everything I’ve built: from a customer-facing Next.js platform tuned to a 97 Lighthouse score, to a compiler overhaul that took aerospace simulations from ten days to three and a half hours.

Now pursuing an MS in Computer Science at NYU, I’m working to bridge the gap between the chaotic nature of AI and the rigid reliability of systems engineering — using Nix and reproducible tooling to bring transparency and consistency to AI environments.

15×
Compiler speedup — aerospace simulations from 10 days to 3.5 hours
97
Lighthouse score, up from 58, on a site serving 20k+ monthly visitors
−45%
LLM inference cost via semantic caching and model routing
50k+
Enterprise documents made queryable with grounded, cited RAG
02

Experience

  1. 2026

    AI Engineer Tata Consultancy Services

    • Built a RAG platform on LangChain, Azure OpenAI and pgvector letting enterprise clients query 50,000+ internal documents with cited, grounded answers.
    • Designed agentic workflows with tool calling and structured outputs that resolve 35% of IT service-desk tickets without a human.
    • Stood up an LLM evaluation pipeline across 500+ prompts, cutting hallucination rates by 40%; reduced inference cost 45% with semantic caching and model routing.

    LangChain · Azure OpenAI · pgvector · FastAPI · Kubernetes · OpenTelemetry

    Apr — Sep
    Chennai
  2. 2025

    Software Engineer Yoho

    • Architected the company’s customer-facing website from the ground up in Next.js, TypeScript and Tailwind, serving 20,000+ monthly visitors.
    • Raised Lighthouse performance from 58 to 97 through SSR, image optimisation, code splitting and edge caching — bounce rate down 25%.
    • Shipped a headless CMS and Node/PostgreSQL APIs that cut content turnaround from days to minutes; technical SEO lifted organic traffic 60% in four months.

    Next.js · TypeScript · Tailwind · Node.js · PostgreSQL · CI/CD

    Sep ’25 — Apr ’26
    Chennai
  3. 2025

    Software Engineer Intern Modelon

    • Led optimisation of the OptiMica compiler and integrated LLMs into the cloud platform, defining efficiency-versus-accuracy criteria for probabilistic models.
    • Delivered agentic automated testing covering 250+ existing cases in GitLab CI; cut suite build and run time by 30%.
    • Migrated CI from Jenkins to GitLab, reducing overhead cost by 60%, and grew reusable k3s/k8s modules by 20%.

    Compilers · GitLab CI · Docker · BuildKit · k8s

    Jun — Sep
    Tiruchirappalli
  4. 2024

    Software Engineer Intern 1Stop.ai

    • Engineered AWS architecture on Elastic Beanstalk and EC2 for deep-learning inference across five model endpoints, with auto-scaling for 10,000+ monthly requests.
    • Reduced inference latency by 20% and reached 99.9% model-serving uptime with an MLOps pipeline and CI/CD for rapid model updates.

    AWS · EC2 · RDS · Azure · MLOps

    Aug — Nov
    Chennai
  5. 2024

    Software Engineer Intern Venkateswara Agency

    • Built a Python desktop billing system over SQLite managing 10,000+ records, with digitally signed tamper-proof invoices and a weighbridge API integration — halving manual effort.

    Python · SQLite · PyHanko

    May — Aug
    Madurai
  6. 2023

    Software Engineer Intern Retech Solutions

    • Automated data-processing workflows in Python and built a full-stack Flask application.

    Python · Flask

    Jun — Jul
    Chennai
03

Research

  1. 2024

    Research Assistant, Optimisation & Compilers Amrita University

    • Overhauled the Modelica toolchain at the compiler level for a 15× speedup, taking complex aerospace simulations from 10 days to 3.5 hours.
    • Introduced function inlining, scalarisation, array expansion and parameter propagation, cutting runtime of previously infeasible models by 70% on average.
    • Built five custom compilation passes for aeronautical workloads, halving the compute cost of three high-fidelity simulation procedures.

    Modelica · Compiler Optimisation · Performance Engineering

    May — Aug
    Coimbatore
04

Selected Projects

P—012025

Efficient CNN for Skin Cancer Detection

A MobileNet-V3–inspired model built for the diverse skin tones of the Indian populace: 95.2% accuracy on 12,000+ images, F1-score disparity across Fitzpatrick types held to 5% (baseline 25%), and 96.7% fewer parameters — 75 ms per image on a Raspberry Pi 5.

Python · CNN · Edge ML

P—022024 — Now

Darwin Configuration

A semi-ephemeral, deterministic macOS system built with nix-darwin and impermanence, with SOPS-encrypted secrets for 10+ cloud providers. Machine setup time down 96%.

Nix · nix-darwin · SOPS

P—032022

Nix Forge

A unified cross-compilation system across Linux, Darwin, WSL and BSD targeting ARM, x86 and RISC-V. 100+ packages from a single source configuration, 87.9% first-try cross-build success, 45% faster foreign-arch CI.

Nix · Cross-compilation · CI

P—042022

JIT Compilation Emulator

An LLVM-backed just-in-time emulator for legacy Java ME games, running at 60 FPS where interpretation stalled, with a reproducible Nix toolchain that turns hours of setup into minutes.

LLVM · JIT · Nix

P—052023 — Now

Blog with Ease

An open-source, Markdown-based blogging system with Giscus comments and GitHub Actions deploys — 80% less manual publishing work.

TypeScript · GitHub Actions · Docker

P—062022

Universal Terminal Previewer

A multi-format terminal file viewer for 100+ file types with rich graphics over the Kitty image protocol, composable with Unix pipelines.

Bash · Kitty Protocol · TTY

05

Capabilities

Languages
TypeScript
JavaScript
Python
Java
C / C++
SQL
Bash
Web
Next.js
React
Node.js
Tailwind CSS
FastAPI
PostgreSQL
Headless CMS
AI / ML
LangChain
RAG
Azure OpenAI
pgvector
LLM Evaluation
Agentic Workflows
Infrastructure
Nix
Docker
Kubernetes
AWS · Azure · GCP
GitHub Actions
GitLab CI
06

Education

  1. 2026

    New York University Tandon School of Engineering

    Master of Science, Computer Science

    Sep ’26 — May ’28
    New York
  2. 2021

    Anna University MNM Jain Engineering College

    Bachelor of Engineering, Computer Science & Engineering

    Aug ’21 — May ’25
    Chennai

Leadership

  1. 2024

    IT Team Lead Cognit’24 Symposium

    Led a 15-person team across four groups to ship the symposium site, a registration system for 1,000+ attendees, and a coding-challenge platform for 500 participants.

    Jan — Mar
    Chennai
  2. 2023

    Event Head Cognit’23 Symposium

    Ran a treasure hunt for 100+ students with 95% positive feedback — recognised as the most efficiently managed event.

    Feb — Mar
    Chennai
  3. 2022

    Committee Member MNMJEC Red Cross

    Led health-awareness campaigns across 20 schools, reaching 500+ students.

    Aug — Oct
    Chennai
07

Contact

Open to software engineering roles and internships in New York and beyond. The fastest way to reach me is email.

contact@surya.sh