Data and AI engineer with 10 years of experience, currently embedded with a US enterprise client to take AI use cases from loosely defined business problems to working prototypes and then into production on the client's …
Data and AI engineer with 10 years of experience, currently embedded with a US enterprise client to take AI use cases from loosely defined business problems to working prototypes and then into production on the client's own platform. Builds end to end: data pipelines, semantic layers, APIs and front ends across Databricks, Snowflake, AWS and GCP. Owns discovery, demos and stakeholder alignment, plus the security, governance and CI/CD work needed to ship in locked-down enterprise environments. Has served pharma, financial services, media and e-commerce clients; technical lead and mentor working with teams in the US, Canada and India.
Experience
S
Principal Consultant, Data Analytics
Systems Limited·Islamabad
May 2024 - Present · 2 years 6 months
<p>Worked directly with client business teams to turn a loosely defined use case, pain points and sample data (no formal requirements) into a unified AI data concierge: dashboards plus a conversational assistant across two brands.</p><p> Built and demoed the prototype end to end in the client's Databricks dev workspace on synthetic data generated from business samples, which led the business to request a move to real data.</p><p> Owned the full stack as four Databricks Asset Bundles: a Lakeflow pipeline (33 sources), two Genie spaces (15 metric views), and a FastAPI backend and React front end on Databricks Apps.</p><p> Led the move to staging on real data, working with source owners and the client platform team to identify unknown systems across Redshift, Snowflake and Unity Catalog and secure access.</p><p> Re-engineered ingestion that broke at full volume into an incremental landing layer over federated sources with billions of rows: watermarks, idempotent MERGE, PII hashing, and Auto CDC.</p><p> Ran a full-volume audit (159 issues, 28 affecting live KPIs), reconciled it with business KPI definitions, and redesigned the pipeline and semantic layer.</p><p> Adapted delivery to client platform constraints, moving from CLI to GitLab CI/CD through the client's shared deployment controller on locked-down runners.</p><p> Drove decisions with client stakeholders on PII handling, data classification and LLM use; added LLM chat summaries via ai_query after comparing models on quality and cost.</p><p> Tested Genie against 140 real business questions and validated migrated data against Redshift; now leading production readiness and rollout.</p>
S
Senior Data Engineer
S&P Global Market Intelligence·Islamabad
Jul 2021 - May 2024 · 2 years 10 months
<p>Designed a reusable MS SQL Server-to-PostgreSQL solution computing change-data-capture via PySpark merge on AWS EMR Serverless, delivering to multiple client systems and channels.</p><p> Built reusable Terraform for GCP (BigQuery, datasets, service accounts, secrets, roles, Dataplex tags) for a data-fabric initiative whose migration roadmap other teams followed.</p><p> Built Spark/Scala pipelines on Databricks and EMR, orchestrated with Databricks Workflows, AWS Step Functions, and EventBridge.</p><p> Ran technology evaluations and R&D to guide the team toward the best-fit tools for business needs.</p>
S
Data Engineer
S&P Global Market Intelligence·Islamabad
Jul 2021 - May 2023 · 1 year 10 months
<p>Built complex Spark/Scala transformations on Databricks and Delta Lake that pivoted and aggregated long-format source data into a wide consolidated target table of up to 1,500 columns, computing deltas against the existing target before load.</p><p> Led the migration of the team's core services from on-prem servers to AWS EC2 as one of the first teams company-wide, eliminating on-prem maintenance overhead.</p><p> Built Kibana monitoring dashboards that cut pipeline-failure detection time and improved observability.</p>
S
Assistant Manager, IT Infrastructure
SUPARCO (National Space Agency)·Lahore
Feb 2020 - Jul 2021 · 1 year 4 months
<p>Administered servers, secure data movement and network security at a satellite control facility.</p><p>Learned about satellite control operations.</p>
<p>Led a team of 8 building an AI-enabled transportation asset system that detected and geo-located roadway assets with TensorFlow.</p><p> Deployed models on AWS SageMaker and served inference through AWS Lambda, storing results in S3 for a customer-facing web application.</p><p> Won first prize, R&D category, P@SHA ICT Awards 2019, for AI-enabled automated traffic-sign identification.</p>
<p>Developed full-stack e-commerce web applications in ASP.NET MVC, Web API, C#, JavaScript and MS SQL, and authored scope-of-work documents from client requirements.</p>
<p>Description: An AI-powered analytics platform on Databricks that unifies about 33 business data sources for two video brands, Sling and DISH. It gives executives dashboards and a natural-language chat interface for KPIs such as viewership, churn, billing, customer experience and NPS/OSAT. Questions are answered with governed, explainable SQL.</p><p>Responsibilities and contributions:</p><ul><li>Built a bronze, silver and gold Lakeflow pipeline in PySpark and SQL. It ingests data from federated Redshift/Snowflake and Unity Catalog sources and standardizes it into brand-specific, Genie-ready models.</li><li>Developed and tuned Databricks Genie spaces for Sling and DISH, including curated example SQL, join specs, instructions and benchmarks. This lets business users get answers in plain English.</li><li>Contributed to a FastAPI backend and a React/TypeScript frontend. They are deployed as Databricks Apps and serve dashboard panels and chat through a shared LLM chat gateway with guardrails and streaming responses.</li><li>Set up CI/CD with Databricks Asset Bundles and GitLab. Dev, staging and production deploy through a controlled pipeline.</li><li>Ran a full data-quality audit. It found about 190 issues, including 3x duplicated source data, silently dropped records and incorrect KPI logic. I fixed them in the silver and gold layers and measured Genie answer accuracy before and after.</li><li>Worked with business stakeholders to turn KPI definitions and formulas into consistent metrics across dashboards and chat.</li></ul><p>Tech stack: Databricks (Unity Catalog, Lakeflow/DLT, SQL Warehouses, Genie, Databricks Apps, Vector Search), PySpark, SQL, Python (FastAPI), React, TypeScript, Databricks Asset Bundles, GitLab CI/CD, LLM/RAG patterns.</p><p></p>
