gowtham. contact ↗

Staff Software Engineer · AI Runtime Security

Gowtham.

I work on the layer between people and their LLMs — catching prompt injections, leaks and toxic content in real time, on the way in and on the way out.

SCROLL TO SEND
POST /v1/chat/completions
{
  "from": "portfolio-visitor",
  "messages": [{
    "role": "user",
    "content": "Tell me about Gowtham"
  }]
}
REQUEST
model.generate()
● INTERCEPTED · BEFORE THE MODEL

→ inbound · request← outbound · response

Every message gets read twice.

POST /v1/chat/completions200 OK · model response
{
  "model": "your-llm",
  "messages": [{
    "role": "user",
    "content": "Tell me about Gowtham"
  }]
}
{
  "role": "assistant",
  "content": "Gowtham is a Staff SWE
    on AI Runtime Security at
    Palo Alto Networks, building
    the platform that serves its
    detection models…"
}

// whoami

~/gowtham

$ cat profile.yaml

name: Gowtham Gopalakrishnan

role: Staff Software Engineer — AI Runtime Security

company: Palo Alto Networks

focus: cloud-native ML infrastructure · AI security · LLM inference

summary: 2+ years building production systems across model serving,

document intelligence and enterprise automation

based_in: Santa Clara, CA

origin: India

$

// what I build

AI security infrastructure, at production scale

I build and run the infrastructure behind AI runtime security — the layer that screens traffic between applications and large language models in real time. My work spans model serving and inference, CI/CD and automated testing, multi-region Kubernetes deployments, and the observability and on-call practices that keep production ML services fast and reliable. The goal: security checks thorough enough to trust and fast enough that nobody notices them.

ML servingKubernetesCI/CDObservability

// experience

HOP 01 · Mar 2026 – Present

Staff Software Engineer — AI Runtime Security

Palo Alto Networks · Santa Clara, CA

  • Build cloud-native ML serving infrastructure for AI runtime security, the services that screen LLM traffic for threats in real time.
  • Work across inference, CI/CD, automated testing and observability on Kubernetes in a multi-region production environment.
  • Share on-call and incident response for production ML services.
PythonKubernetesGCPGitLab CI/CD

HOP 02 · Sep 2025 – Dec 2025

Machine Learning Engineer Intern

OneData Software Solutions · Fort Mill, SC

  • Built a source-agnostic ingestion and vector-indexing pipeline that turned private GitHub and OneDrive documents into searchable embeddings for LLM retrieval.
  • Implemented the end-to-end flow (ingest → extract → chunk → embed → index) with Python, OpenAI embeddings and PostgreSQL + pgvector.
  • Designed deterministic chunking, metadata enrichment and batched vector inserts to scale indexing to thousands of chunks.
~80% less manual document lookup

HOP 03 · Jul 2022 – Apr 2023

Associate Software Engineer (Full-time)

Antworks · Chennai, India

  • Developed features for the company's Intelligent Document Processing platform in Python.
  • Built event-driven document processing flows with Kafka and wrote complex SQL powering classification, entity recognition and reporting.
70% → 90% extraction accuracy−25% processing time+30% throughput

HOP 04 · Feb 2022 – Jun 2022

Cognitive Software Development Intern

Antworks · Chennai, India

  • Built and configured client automation workflows on QueenBot, the company's RPA/automation platform, turning business requirements into production workflow bots.
  • Partnered with the support team to debug and resolve client workflow issues, writing SQL for data validation, investigation and reporting.

// education

Jan 2024 – Dec 2025

MS in Data Science

University of Arizona · Tucson, AZ

Sep 2018 – Apr 2022

BTech in Information Technology

Sri Krishna College of Engineering and Technology · India

// side quests

Built on my own time.

01

Clinical Text Classification with BERT

PyTorch LightningBioBERTRoBERTaOptunaFastAPIDockerW&B

Fine-tuned BioBERT and RoBERTa on MIMIC-III clinical notes with Optuna hyperparameter search and GPU training. Served the best model through a Dockerized FastAPI API, with experiments tracked in Weights & Biases.

83%test accuracy · measured alongside F1 and AUROC for precision–recall balance
GitHub ↗
02

Environmental Impact Scoring (EIS) Platform

PythonSpark StreamingOpenAI APIscikit-learn

A data-driven decision system for sustainability analytics. Streaming pipelines and ML clustering analyze large, continuously updated datasets. Interpretable visuals (cluster maps, correlations, trends) support automated insights and human-in-the-loop decisions.

Streaming MLreal-time clustering
GitHub ↗

// sandbox

Try to get past the hand.

A tiny keyword toy standing in for the real thing — production detectors are ML models, not regexes. Try a jailbreak, leak a fake secret, or just say hi.

// capabilities.yaml

What's loaded.

core_stack:PythonKubernetesGCPTritonvLLMGitLab CI/CD

languages:

3
PythonSQLBash

ml_serving_and_ai_infra:

7
Triton Inference ServervLLMVertex AIPyTorchLLM inference & guardrailsRAGAI-powered workflows

cloud_gcp:

6
GKECloud RunCloud Storage (GCS)BigQueryArtifact RegistryIAM & service accounts

containers_and_iac:

5
KubernetesDockerHelmTerraformmulti-region deployment

ci_cd_and_testing:

5
GitLab CI/CDpytestDocker-based e2e & smoke testingpre-commit hooksgo-task

observability_and_reliability:

5
PrometheusGrafanaLokistructured loggingon-call & incident management

backend_and_data:

7
FastAPIgRPCPydanticKafkaRedisPostgreSQL / pgvectorMySQL

// prior interceptions

Catching bad actors, before it was my job.

Three of these four were security and law-enforcement challenges.

ID · 2023

Karnataka State Police Hackathon

Karnataka State Police, India

WINNER

ID · 2022

DevBhoomi Cyber Hackathon

India

WINNER

ID · 2022

Darkathon

Narcotics Control Bureau, India

RUNNER-UP

ID · 2024

AftroTech AI Hackathon

USA

RUNNER-UP

// 200 OK · response delivered

Let's build safer AI.

Open to conversations about AI security, model serving, and ML infrastructure.

gowthamgopalakrishnan126@gmail.com · Santa Clara, CA

Email copied