Skip to content
View veerjain-1's full-sized avatar

Block or report veerjain-1

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
veerjain-1/README.md

Hi, I'm Veer Jain 👋

Distributed Systems • AI Systems & Security • High-Throughput Pipelines

Portfolio LinkedIn Email


Computer Science & Artificial Intelligence (Double Major) @ Purdue University
Ex-AWS • Ex-Lockheed Martin • Ex-Textron • Ex-John Deere


🚀 About Me

I'm a systems engineer and builder focused on distributed systems, ML infrastructure, and high-throughput data pipelines. I enjoy taking complex engineering challenges down to their first principles—whether that's serving real-time CodeBERT inference through a Rust gRPC engine, building distributed event-driven pipelines over Apache Kafka, or designing sub-millisecond deterministic security & reliability evaluation frameworks for autonomous AI agents.

  • 🔭 Current Focus: Distributed systems, low-latency microservices, ML inference infrastructure, scalable backend architectures, and AI agent evaluation & security.
  • 💡 Engineering Philosophy: Building high-performance, resilient software that scales horizontally, minimizes latency overhead, and operates reliably under heavy production workloads.

🌐 Open Source Contributions

I actively contribute upstream to foundational machine learning frameworks, distributed workflow engines, and developer tooling:

Project PR Technical Contribution Domain / Subsystem
tensorflow/tensorflow #125400 Added bounds check on seq_len_max in BlockLSTM ops to prevent out-of-bounds memory access Deep Learning C++ Kernel Safety
microsoft/semantic-kernel #14284 Implemented Vertex AI :batchEmbedContents routing for gemini-embedding models LLM Embedding Infrastructure
cadence-workflow/cadence #8525 Replaced fixed sleeps with concurrent domain notification polling in integration tests Distributed Orchestration CI Engine
run-llama/llama_index #23127 Parallelized recursive retrieval using concurrent run_jobs while preserving exact rank order Query Retrieval Latency Optimization
microsoft/vscode-containers #598 Added Docker Compose pull support across palettes, context menus, and tree views Cloud Developer Environments
kirodotdev/KiroCrew #3146 Bounded interpreter probe with POSIX watchdog to prevent hangs on macOS CLI Tooling & Process Safety

🌟 Featured Systems & Architectures

Real-time Code Anomaly & Vulnerability Detection Platform

  • Polyglot event-driven microservices architecture in Go + Rust + Kafka + React.
  • Sub-50ms ML inference via custom-trained CodeBERT exported with dynamic axes running on ONNX Runtime in Rust over gRPC.
  • Buffered through Apache Kafka (KRaft mode) with live WebSocket streaming to an analytics dashboard.

Image Image Image Image Image

Deterministic Security, Guardrail & Eval Framework for AI Agents

  • Sub-millisecond trace evaluation engine eliminating flaky and costly "LLM-as-a-judge" overhead.
  • Deterministic rules detecting infinite execution loops, schema drift, command injection, path traversal, and credential leakage.
  • Adversarial prompt injection fuzzer and native SARIF v2.1.0 export for GitHub Security alerts.

Image Image Image Image

Autonomous CI/CD Workflow & Code Review Controller

  • Orchestrates multi-stage build workflows and real-time dependency validation using Java (Spring Boot) and LangGraph.
  • Asynchronous Kafka message queues trigger recursive LLM agents for deep-dive code analysis and comprehensive PR summaries.
  • Stateful workflow observability with full Prometheus & Grafana instrumentation.

Image Image Image Image

Domain-Specific Financial SLM with SFT & RAG

  • Supervised fine-tuning (SFT) pipeline on financial datasets utilizing PyTorch and Hugging Face Transformers.
  • Leverages Apple Metal Performance Shaders (MPS) for 7x–10x local training acceleration over CPU.
  • Domain-adapted retrieval-augmented generation (RAG) with FAISS vector indexing for high-precision financial QA.

Image Image Image Image


💻 Technical Arsenal

Category Technologies
Languages Go • Rust • Python • Java • C++ • C • TypeScript • JavaScript • SQL
Distributed & Cloud Apache Kafka • AWS (Bedrock, Fargate, CDK, DynamoDB) • Docker • Kubernetes • gRPC / Protobuf • Redis
ML Infrastructure & Evals PyTorch • ONNX Runtime • CodeBERT • Hugging Face • LangGraph • Deterministic Evals • Red-Teaming
Databases & Storage PostgreSQL • MongoDB • Apache Spark • FAISS • MySQL
Frontend & UI React • Tailwind CSS • Framer Motion • Recharts • Vite

Built with care by Veer Jain • Explore more on my Portfolio

Pinned Loading

  1. code-anomaly-engine code-anomaly-engine Public

    Real-time code anomaly detection engine — Go + Rust + Python + React. Custom-trained CodeBERT model with sub-50ms inference via ONNX Runtime.

    Go 2

  2. agent-guard-eval agent-guard-eval Public

    Deterministic security, guardrail & reliability evaluation framework for AI agents with JUnit and SARIF reporting

    Python 2

  3. agentic-cicd-orchestrator agentic-cicd-orchestrator Public

    Agentic CI/CD Orchestrator using Java, Spring Boot, LangGraph4j, Kafka, and Prometheus.

    Java 1

  4. ai-code-review-platform ai-code-review-platform Public

    High-scale microservices backend for automated AI code reviews — Node.js + Java + Kafka + LangChain

    JavaScript

  5. finance-slm-rag finance-slm-rag Public

    Domain-specific SLM for finance with RAG and PyTorch SFT

    Python 1

  6. veerjain-1.github.io veerjain-1.github.io Public

    Personal portfolio & interactive AI agent — React, Tailwind CSS, Framer Motion, Vite

    JavaScript