AI / ML Engineer · Philadelphia, PA
Kiran
Shidruk

Building intelligent systems at the intersection of machine learning, real-time pipelines, and agentic AI. MS in AI/ML from Drexel University — currently leading technical initiatives at ShipCube.

View Experience GitHub Contact
Kiran Shidruk
Open to Opportunities
6+
Years Experience
100K+
Daily Transactions
3.93
MS GPA
10M+
Daily Logs Processed
01

About Me

I'm an AI/ML Engineer with 6+ years of experience building production-grade machine learning systems across cybersecurity, HR tech, and supply chain domains.

At ShipCube, I lead technical initiatives spanning forecasting models, context-aware AI features, agentic automation, and real-time data pipelines — managing a team of 8 engineers and serving 100+ operational stakeholders.

I hold an MS in AI & Machine Learning from Drexel University (GPA: 3.93) and have research experience in Vision-Language Models. I believe the best ML work is invisible — it just makes things work better.

Areas of Focus
01
ML / DL Model DevelopmentXGBoost, LSTMs, Transformers, ensemble methods — end-to-end from feature engineering to production inference.
02
Agentic AI & AutomationBuilding autonomous workflows with LLMs and orchestration tools to eliminate manual processes at scale.
03
NLP & Computer VisionVLMs, zero-shot classification, multimodal learning, and applied NLP pipeline development.
02

Skills

Programming & Databases
PythonJavaJavaScriptC++GoLangMySQLMongoDBChromaDBPineconeSystem DesignDSA
AI / ML
XGBoostProphetLSTMsTransformersNLPComputer VisionLLMsAgentic AIAnomaly DetectionVector SearchA/B Testing
Cloud & MLOps
GCPVertex AIBigQuery MLCloud RunPub/SubDataprocAWSAzureDockerKubernetesCI/CDMLOps
03

Experience

ShipCube
Jan 2026 – Present
Early Stage Startup
Current
AI/ML Engineer
  • Engineered ensemble forecasting pipeline (XGBoost, Prophet, LSTMs) for sales & inventory demand, cutting stockouts by ~35% and excess order volume by ~40% via automated demand smoothing.
  • Built context-aware AI recommendation engine (Vertex AI, BigQuery ML, vector search) for personalized reorder suggestions and data-driven warehouse location assignments; boosted customer engagement by 40%.
  • Led 8-engineer team to architect event-driven order management platform (Cloud Run, Pub/Sub, BigQuery) processing 100K+ daily transactions at sub-200ms p95 latency and 99.95% uptime.
  • Designed agentic AI workflows on n8n with internal API integrations to automate Finance, Sales, and Operations processes — eliminating manual overhead across departments.
  • Delivered real-time dashboards for 100+ stakeholders surfacing live order, inventory, and fulfillment metrics.
  • Implemented CI/CD and observability stack (Cloud Monitoring, Cloud Trace, Cloud Build), reducing deployment time by 70% and MTTR by 55%.
Drexel University
Jan 2024 – Mar 2025
Philadelphia, PA
ML Research Intern
  • Led research on Vision-Language Models (GPT-4V, LLaVA, Gemini Pro, BLIP-3) for zero-shot scientific diagram classification; ran 20+ experiments across 50+ multimodal papers, improving accuracy by 10% and robustness by 18%.
Great Software Lab
May 2021 – Dec 2023
Fast-paced Product Team
Data Scientist
  • Transitioned rule-based threat detection to ML-driven pipeline, developing Naïve Bayes and deep learning anomaly detection models on HDFS to identify abnormal user/device behavior across millions of daily logs (Splunk, QRadar, Datalake).
  • Engineered dynamic risk scoring engine assigning real-time risk scores based on event-triggered model inference, feeding downstream security monitoring and incident response applications.
  • Led GCP migration from HDFS, architecting scalable pipelines (Pub/Sub, BigQuery, Spanner, Cloud Storage) enabling real-time log processing — reducing detection latency over legacy batch workflows.
Hefshine Softwares
Nov 2019 – May 2021
Early-stage Startup
ML Engineer
  • Developed Random Forest–based employee attrition model (AUC-ROC 0.78) via Flask–Spring Boot API, enabling real-time risk scoring and reducing turnover by 10%.
  • Implemented K-Means segmentation and Power BI/Tableau dashboards, improving appraisal efficiency by 20%; architected ML inference pipeline with <500ms latency.
04

Projects

i.
Multilingual AI Legal Assistance
5th / 100+ teams at Philly Codefest 2024. AI app providing accessible legal guidance across language barriers using LLMs and multilingual NLP pipelines.
LLMsNLPMultilingual
View on GitHub →
ii.
Vision-Language Model Research
Zero-shot scientific diagram classification using GPT-4V, LLaVA, Gemini Pro, BLIP-3. 20+ experiments — improving accuracy by 10% and robustness by 18%.
VLMsZero-ShotPyTorch
View on GitHub →
iii.
Supply Chain Forecasting Engine
Ensemble system combining XGBoost, Prophet, and LSTMs for demand prediction. Reduced stockouts by ~35% and excess order volume by ~40% at ShipCube.
XGBoostProphetLSTMs
View on GitHub →
iv.
Threat Detection ML Pipeline
Migrated SIEM rule-based system to ML-driven anomaly detection with Naïve Bayes and deep learning, processing millions of daily logs from Splunk and QRadar.
Anomaly DetectionHDFSGCP
View on GitHub →
05

Writing

Agentic AI
Building Agentic Workflows with n8n and LLMs Soon
2026
Forecasting
Ensemble Forecasting for Supply Chain: XGBoost + Prophet + LSTMs Soon
2026
MLOps
From HDFS Batch to GCP Streaming: A Migration Story Soon
2025
Research
Zero-Shot Classification with Vision-Language Models Soon
2025
06

Education

MS in Artificial Intelligence & Machine Learning
Drexel University
College of Computing & Informatics · Sep 2025 · Philadelphia, PA
GPA 3.93 / 4.0
BS in Electronics Engineering
Mumbai University
Jun 2019 · Mumbai, India
GPA 8.1 / 10.0
07

Awards

2024
5th / 100+ Teams — Philly Codefest, Multilingual AI Legal Assistance App
2023
MIT-WPU PG Program in Artificial Intelligence & Machine Learning
2022
Top 3 / 250 — Pat on the Back Award, Excellence in Scalable ML System Design
2022
IBM Professional Data Science Certification
Let's
Connect

Open to AI/ML Engineer roles and interesting collaborations. Feel free to reach out via any channel below.

Emailkiran.shidruk.us@gmail.com LinkedInlinkedin.com/in/kiranshidruk GitHubgithub.com/kiranshidruk Phone+1 215 240-9822