Engineering Director | "People" first - not "resources" | Drive Large Scale Distributed Platforms | AI/ML - DeepLearning | Data Engineering/Big Data | GCP Architect | Financial Risk | Learner

Seasoned Engineering Director, currently driving Digital Acquisitions at American Express, with more than 20+ years of experience leading large-scale applications in fintechs. Proven track record in driving over $40 million tech investments per year and managing products impacting 20+ global markets, 10M+ users and 10,000+ servicing center reps. Expertise in AI/ML, GenAI, Data Engineering/Analytics and financial technologies (credit and fraud risk) using Java/Python/Go. Good experience in LLM training, data curation, and fine-tuning foundation models like Llama. Seeking engineering leadership
roles in driving exponential business growth.

Skills
• Technical Leadership Technical roadmap for AI/ML and GenAI and delivery of Large-Scale Distributed Applications
• Strategic Planning Partnering & Collaborating with C-Level Executives, Researchers for technical and product strategy.
• People Leadership Exemplar of organization culture, Building and mentoring cohesive global distributed teams.
• Influencer Managing key external customer technical relationships including Amazon, Lowes, Delta, Hilton.

✉ shail.bade@gmail.com

Shailendra Bade

AI agents are no longer living in laboratories. 
They're out in the wild, making decisions, writing code, and connecting with everything in sight. 
And guess what? 
The old security rules don't work anymore. ⚠️

Security baked in, not bolted on. 
We've all heard it. 
But now, with autonomous AI agents, it's not just a nice idea. 
It's survival. 🔒

Let's talk reality: Meta's LlamaFirewall is a wake-up call. 
It's not just another firewall. 
It's a signal that attackers are catching up to AI, and our playbook needs an overhaul. 🚨

Here's how the attack surface has changed: 
- Jailbreaking 2.0: Hackers don't just tweak outputs, they hijack agent goals. 
- Goal hijacking: Feed the agent the wrong input and you can reroute its entire mission. 
- Code injection at scale: If your AI writes code, every output is a security risk.

I've run my own POCs using layered defenses, inspired by this new approach. 
Real numbers, real results: 
- Input filtering blocked threats with over 97% accuracy. 
- Behavioral monitoring using lightweight models caught agents straying from intended goals. 
- Output validation checked every line before execution. ✅

The outcome? 
- Successful attacks dropped from nearly 18% to under 2% in controlled tests. 
- Small, efficient models (86M parameters) outperformed their heavyweight cousins for security tasks. 
- False alarms stayed below 2.5% and didn't slow down the systems.

Here's the uncomfortable truth: 
If your AI agent can search the web, write code, and act on its own, the entire system becomes an attack surface. 
The security fence moves with every new feature you ship.

The organizations that win in the autonomous AI era? 
They're the ones building security into the DNA of their AI from the very beginning. 
Not via quick patches. 
Not by "adding a firewall later." 
But by rethinking how AI is designed, deployed, and monitored. 🛡️

Leaders, the time for afterthought security is over. 
If your plan starts with "we'll add protection later," you're already behind.

Engineering Director | Gen AI/RAG |AI/ML & LLM Systems at scale | From Financial Fraud Detection to Personalization | Data Engineering & Privacy

XtractPro Software

•	Since Jan 2024, I create successful video courses on Udemy and other websites.
•	Designed and implemented end-to-end ML pipelines, with Mlflow, Streamlit, Snowpark… Saved customers $100K+/y overall by switching to some open-source products.
•	Designed data lake solutions with Snowflake, Databricks, on S3 and other cloud storage… Saved one single customer over $30K/y just by avoiding data silos and moving data. Check my open-source Snowflake productivity tools on GitHub (https://github.com/cristiscu?tab=repositories), described on my blog (https://cristian-70480.medium.com/).
•	MLOps model deployment and CI/CD, with custom data drift monitor, TF Serving, Kubernetes…
•	MLOps and Model Ops for batch inference, with AWS Lambda functions, and REST APIs. The customer was wasting money for 2y+ by not having a proper ML pipeline for prod deployment.
•	Implemented an anomaly detection, two recommender systems, time series predictions…
•	EDA and feature engineering on multiple datasets, for ML with sklearn and TensorFlow/Keras.
•	Designed and built feature stores for Snowflake and Redis, with Feast and AWS. Serving key-value Redis data much faster for real-time production ML inference was a huge bottleneck until now.
•	Advised customers on Explainable AI and Ethical AI (with Diff Privacy and Model Fairness demos).
•	Implemented many ML prototypes and PoCs with AutoML and specialized CV/NLP cloud services.

Independent Consultant and Instructor

Onica by Rackspace Technology

•	Lead the adoption and delivery of the Snowflake practice.
•	Snowflake Subject Matter Expert (SME) with the pre-sales, sales and delivery teams.
•	Developed and held presentations with all Snowflake-centric architectures and ecosystems.
•	Mentored peers and less senior architects and engineers.
•	Worked with the new Data Science and Machine Learning groups to extend the Snowflake offer.
•	Helped develop and expand the Snowflake product offer to potential customers.
•	Analyzed and designed several possible solutions for real-time data streaming to Snowflake.
•	Designed PoCs of Snowflake ecosystems in AWS/Azure/GCP.
•	Technical lead on several AWS/Azure/GCP customer projects.

Lead Data & Solutions Architect (Snowflake SME)

Infostrux Solutions

•	Lead tech resource on a long-term contract with LoyaltyOne/AirMiles, renewed several times because of me (billed 30h+/week). Managed Agile/Scrum meetings with 3-15 people.
•	Designed a new Snowflake/AWS-based ETL/ELT pipeline for LoyaltyOne partners. JS stored procs and AWS lambdas for ingestion. BDD Python integration tests.
•	Consulted on Snowflake analytics/performance with Tableau/Looker. Data clean rooms, auto-scaling, RBAC, access policies. Published posts on Medium on Snowflake's blog.
•	Airbyte vs Fivetran eval. CDC Postgres to Snowflake ingestion, with Kafka/Debezium and dbt.
•	Complex PoCs with Amazon Managed Workflows for Apache Airflow, using EKS/ECS, Docker.
•	Complex SQL for semi-structured JSON in Snowflake data lakes.

Sr. Data & Solutions Architect (Snowflake "Data Superhero")

•	Advised over two dozen customers on cloud migration and data pipeline design, using partitioning, replication, failover, auto-scale (see https://data-xtractor.com/nosql-consulting-services/).
•	Designed software and solution architectures, using N-Tier ASP.NET end-to-end systems, pure JS MEAN stacks, batch processing, real-time streaming or serverless microservices.
•	Advised customers and designed proof-of-concept for drastically improving data access and latency using a NoSQL solution in DynamoDB, Cassandra, MongoDB, Redis.
•	Created over a dozen ETLs for paying customers for batch processing and real-time data migration from relational to NoSQL stores or data warehouses, using SSIS, BI tools, Azure ML Studio, cloud ETLs, Step Functions, Azure Batch.
•	Paired up with Data Scientists to create ETLs or environments for ML predictions, graph algorithms on Neo4j, Azure Stream Analytics queries, advanced analytics.
•	Designed and implemented parts of several ASP.NET projects, using microservices with Azure Functions and AWS Lambda, RESTful web APIs, in C#, Python, Go, Java, JavaScript front-ends.
•	Created dozens of IaC with Terraform (I'm Certified in Terraform), CloudFormation and Azure Resource Manager, as blueprint to quickly create cloud environments for potential clients.
•	Designed over a dozen data models for customers using RDBMS/DW or NoSQL: DynamoDB, Cassandra, MongoDB, Redis, Neo4j.
•	Designed, implemented and sold my Data Xtractor Suite (https://data-xtractor.com) to dozens of customers. This includes an innovative ER modeler, visual SQL query builder, many data visualization chart types. Got paid for extended customer support and additional jobs.
•	Published Kindle books (https://amazon.com/author/cristian-scutaru) and Udemy courses (https://www.udemy.com/user/cristian-scutaru) on Cassandra/Neo4j/Redis/Couchbase (I'm Certified on all).

IT Management Consultant & Architect

Vision Critical

•	Experienced polyglot programmer in a central Agile/XP team.
•	Designed and implemented a complex ETL/ELT pipeline using window functions (LEAD/LAG cursor) on Amazon Redshift, for ~5GB daily transfer/processing and ~500GB monthly uploads.
•	Designed parts of a MongoDB database holding JSON customer data.
•	Design and implementation of new dataviz and analytics, considering CQRS/Event Sourcing..
•	Helped design, improve and implement a complex distributed platform for community insights, with ASP.NET MVC/RESTful Web API, AngularJS/KnockoutJS, SQL Server/Oracle.

Sr. Big Data & Software Engineer

Metalogix Software

•	Hiring manager for team of 5 devs, conducted hiring interviews and held performance reviews.
•	Certified Scrum Master over the largest team in the company (10-12 members), buffer between devs/QA and the Product Owner, Product Unit Manager, upper management and other stakeholders.
•	Also acted as senior .NET/C# Lead Developer for company's main product, to migrate and synchronize content between different SharePoint versions.

Development Lead & Manager (on SharePoint Apps)

•	Successful consultancy business. Multiple contracts as managing consultant or main technical advisor to CTOs, directors and other executives of Vancouver area and US corporations – including the former Ministry of Indian Affairs Canada (INAC), FDM Software, Seance Software... 
•	Helped implement several successful high-traffic websites with significant transaction volume.
•	Designed, implemented, and sold for profit several personal projects, such as a chart template customizer, a website for birthday party locations, niche websites created for SEO…

Boeing Canada-AeroInfo

•	Helped design, implement and refactor complex reporting and data processing modules of a large Java/C# management system for plane parts.

Sr. Software Engineer

Air Liquide

•	Designed the system architecture and helped with technology choices for a large $10M distributed system, later implemented in C#/ASP.NET. 
•	The only software architect, acting as technical lead over a group of 10-15 developers, one DBA, 2-3 analysts.

Sr. Data & Software Architect

Microsoft Corporation (Redmond HQ)

•	Performance tests on the first OLAP server engine (SSAS) of their first Microsoft SQL Server Data Warehouse, in the SQL Server 7 group. ER and multidimensional data modeling.
•	Designed their new (at that time) Microsoft Repository browser. 
•	Helped design and implement their new Windows Clustering Manager (with load balancing, failover and data recovery, from 2/4 nodes), in the Windows 2000 group.

Sr. Software Developer

•	World-class expert in Snowflake, former Snowflake SME and Snowflake “Data Superhero”.
•	Hands-on Data Architect for data lakes in AWS, Azure and GCP (20+ certification exams).
•	Data warehouse expert in Snowflake, Databricks, Redshift, BigQuery.
•	Skilled in Applied AI services, ML pipelines and MLOps architectures.
•	Former Microsoft employee, in Windows and SQL Server groups. US Patent Pending in 2001.
•	Canadian citizen. Worked in Canada, US, Switzerland. Fluent in English, French, Romanian.

taplio

The Battle

Shailendra Bade

Cristian Scutaru

Followers

Posts

Engagements

Likes

Performance ScoreShailendra Bade's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Shailendra Bade

Cristian Scutaru

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Shailendra Bade's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreShailendra Bade's LinkedIn Performance