Looking for a job

Daksh Goel

@dakshgoel

Gen AI Engineer | Amazon ML Summer School ’26 (Top 2%) | PayPal Career Academy Fellow | Agentic Systems, RAG & Multi-Agent Systems

India

Message

What I'm looking for

I’m looking for a role building agentic AI systems in production—focused on reproducible evals, fast iteration, and strong engineering rigor. I enjoy mixing deterministic guardrails with LLMs and shipping with FastAPI/React and CI-gated testing.

I’m a B.Tech AI & ML student (3rd year, CGPA 8.75) focused on building agentic AI systems—especially multi-agent orchestration with CrewAI/LangChain, local-LLM inference, and guardrails that combine deterministic logic with LLM behavior.

I’ve shipped production-style workflows behind FastAPI + React and built evaluation harnesses that prioritize reproducibility and speed. As a PayPal Career Academy 2.0 fellow (1 of 46 nationally), I prototyped an anomaly-detection module that backtested on a quarter of historical fraud-flag logs, cutting false positives by 15% for manual review.

In cybersecurity-focused work at VOIS, I built an end-to-end LLM security pipeline classifying 5+ threat categories and reduced SOC triage time by 40%. I also built real systems end-to-end, from a Python ETL pipeline reducing weekly processing from 25 to 10 hours to an agentic, evidence-grounded research pipeline and high-accuracy dashboards with FastAPI, Qdrant, Firebase, and OpenCV.

Experience

Work history, roles, and key accomplishments

Tech Trainee

PayPal Career Academy 2.0

Feb 2026 - Jul 2026 (5 months)

Selected as 1 of 46 fellows nationally for PayPal’s flagship Career Academy 2.0. Prototyped an anomaly-detection module for distributed transaction data and backtested on a quarter of historical fraud-flag logs, reducing false positives flagged for manual review by 15%.

Anomaly Detection Fraud Analytics Distributed Transactions Data Evaluation

Cybersecurity Intern

Vodafone Intelligent Solutions (VOIS)

Dec 2025 - Jan 2026 (1 month)

Built an end-to-end LLM security pipeline classifying 5+ threat categories, cutting SOC triage time by 40%. Improved classification accuracy by 35% by redesigning prompt strategies and preprocessing pipelines for security datasets.

LLM Security Prompt Engineering data preprocessing

Software Development Intern

Infotact Solutions

Jun 2025 - Aug 2025 (2 months)

Built a Python ETL pipeline for three internal workflows, reducing weekly processing time from 25 to 10 hours and eliminating manual errors. Integrated four third-party APIs for structured data ingestion, reducing response latency by 30% via connection pooling.

Python ETL API Integration Data Ingestion Connection Pooling