Shahwar Saleem

At EvenUp, we use AI to help personal injury lawyers maximize case values by automating demand letter generation and legal document preparation.

Role: Senior Software Engineer (AI and Docgen)

Impact Created:

• Built advanced DOCX-grade document formatting capabilities (indentation, margin controls, colored highlights, font/table styling) impacting over 200,000 existing documents
• Implemented full header/footer support comparable to Microsoft Word, enabling enterprise-grade document compliance and unlocking new revenue opportunities
• Partnered with product and legal stakeholders to translate document-layout requirements into scalable backend abstractions

At Ripple, we provide enterprise blockchain solutions and cryptocurrency payment networks for global financial institutions.

Role: Software Engineer II (ML and Data Infra)

Impact Created:

• Built MLegobricks, a unified ML engineering mono-repo with automated packaging, CI/CD templates, and semantic versioning—standardizing workflows for 12 ML teams and eliminating redundant code
• Automated ML platform migration from GCP to AWS Databricks, reducing 3-4 month timeline to hours through config-driven Python tooling
• Designed secure inference gateway with OAuth2, JWT, and Envoy, establishing decoupled infrastructure patterns for independent scaling

At Arctic Wolf, we provide security operations as a service, delivering 24/7 monitoring, detection, and response for thousands of enterprises.

Role: Software Engineer II (ML Platform)

Impact Created:

• Led evaluation and rollout of Flyte as unified workflow orchestrator, creating standardized platform for ML pipelines and eliminating custom orchestration code
• Delivered first AWS EKS-based Flyte deployment with configuration-only workflow definitions, reducing team onboarding from weeks to 1-2 days
• Contributed upstream fixes to Flyte open-source project and enabled automatic scaling for production ML workloads

At Borealis AI, we advance machine learning research and apply AI to solve real-world financial services problems.

Role: Software Engineer II (ML and Data Platform)

Impact Created:

• Built Anahit, an internal feature store and ML data serving platform standardizing data infrastructure for 20+ research teams and eliminating thousands of lines of custom ETL code
• Developed Spark-based ETL pipelines processing 30TB+ daily from Teradata EDW with Petastorm-materialized Parquet writes
• Implemented advanced partitioning strategies (bucketing, sorting, date-based) reducing data access from seconds to milliseconds for time-series datasets with millions of records
• Integrated PyTorch dataloaders with multi-worker parallelism and multi-stage transformation pipelines via Hydra YAML DSLs

Data Engineering Nanodegree from Udacity.

Key Learnings:

• Deep dive into data modeling (relational and NoSQL), cloud data warehouses (Redshift, BigQuery), and data lakes
• Designed star/snowflake schemas and built ETL pipelines with Airflow, implementing data quality checks
• Optimized query performance for analytics workloads and gained practical experience with distributed processing frameworks (Spark)

Accelerating Data Engineering Pipelines from NVIDIA.

Key Learnings:

• Learned to accelerate data pipelines using GPU computing with CUDA-accelerated libraries like cuDF (GPU DataFrame operations), cuML (GPU machine learning), and RAPIDS ecosystem
• Explored techniques for parallelizing data transformations and optimizing memory transfers between CPU and GPU
• Achieved 10-100x speedups on large-scale data processing tasks through GPU acceleration

At ProNavigator, we provide AI-powered virtual assistants for customer service, helping organizations automate support and improve customer experience.

Role: Machine Learning Engineer

Impact Created:

• Optimized NLP models achieving 100× speedup on inference and 100× model size reduction through quantization and distillation, significantly reducing infrastructure costs
• Improved classification accuracy from 95% to 97% through model architecture improvements, enabling seamless production transition with positive customer feedback
• Built confusion matrix visualizations for data labelers, improving intent labeling accuracy from 79% to 90%

MEng Computer Science at University of Waterloo. GPA: 89%

Research & Achievements:

• Graduate Research Assistant at Autonomous Vehicles Lab, focused on ML and Software Architecture
• Co-authored: "An Automated Vehicle Safety Concept Based on Runtime Restriction of the Operational Design Domain" at IEEE Intelligent Vehicle Symposium. Designed and implemented a strategy to alert the autonomous vehicle when a camera is obstructed using Deep Neural Networks and ROS
• Designed build systems for ROS components deployed on autonomous vehicles
• Ported ROS to QNX over ARM architecture, enabling real-time operating system integration
• Coursework involved rigorous mathematics and machine learning theory with emphasis on practical software architecture patterns

At Mentor Graphics (now Siemens), we were a giant in electronics and embedded software development, famous for the Nucleus RTOS—a proprietary layered-approach real-time operating system.

Role: Software Engineer

Impact Created:

• Designed and implemented Nucleus Kernel Awareness (UI + Backend) for ReadyStart IDE, enabling OS developers and application programmers to debug by inspecting running instances of OS structures in real-time
• Developed BSP Validation Kit, a Python-based driver test automation platform for Board Support Package validation, standardizing testing workflows across the Nucleus BSP team
• Built RTOS device drivers for LCD Display, I2C, SPI, EEPROM, NAND, and NOR across multiple ARM platforms (imx28evk, atsam9263ek, am335x, Vybrid, Zynq) and PowerPC targets (mpc8313erdb, p1020rdb, p2020ds)
• Contributed to toolchain development for ReadyStart, Mentor's Eclipse-based IDE, improving embedded development workflows for Nucleus RTOS applications

Electrical and Computer Engineering at UET Lahore, Pakistan. GPA: 87%

Key Achievements:

• Comprehensive electrical engineering degree with computer science emphasis, providing deep understanding of how computers work from hardware to software
• Capstone project: Developed a Beowulf Cluster customized for compute loads, gaining hands-on experience with distributed computing and parallel processing architectures