Engineering Portfolio

Misbahuddin Mohammed

Senior Engineering Leader

11 years at Amazon building and scaling engineering organizations across AI/ML, data platforms, and logistics operations. From real-time ML pipelines to LLM-powered data infrastructure — leading teams, shipping products, and driving $45M+ in combined business impact.

Read Resume

By the numbers

Portfolio Impact

11 years of building systems, growing teams, and delivering measurable business outcomes across logistics, AI/ML, and data platforms.

Total portfolio impact

$0

M+

Cost savings and revenue uplift across all products

Engineers grown

3 → 0

Across India, Dubai, and Mexico

Engineers promoted

0

With 12-month evidence-based promo cases

Product deployments

0

regions

IN · AU · SG · DXB · KSA · LATAM

Amazon tables served

0

K+

By AMG metadata platform

Interviews conducted

0

+

Across Amazon hiring initiatives

What I built

Engineering Portfolio

11+ years. 6 products. 3 continents.

2022–Present
AI/ML

DataOps Suite

LLM-powered data catalog, metadata generation, and PII obfuscation platform serving 203K+ internal tables.

$1.2M annual savings · 203K+ tables served

Claude LLM
AWS Lambda
TypeScript
Explore →
2016–2018
Operations

Heimdall (InboundIQ)

ML-driven dock door allocation engine that replaced all manual truck prioritization at fulfillment centers.

TAT 6.7 → 2.2 hours · P95 SLA met consistently

Python
ML Models
AWS Lambda
Explore →
2018–2022
Logistics

Delay Alert Dashboard

Real-time linehaul delay visibility and automated rescue planning across 6 global regions.

6 regions · 10-min refresh · replaced manual spreadsheets

Apache Kafka
React
AWS ECS
Explore →
2016–2018
Fraud Detection

LoFAT

GPS telemetry fraud detection for last-mile delivery — automated flagging of spoofing, ghost deliveries, and coordinated fraud.

$0.6M saved · 37 headcount hires eliminated

ML Anomaly Detection
GPS Telemetry
Python
Explore →
2018–2022
Logistics

Daily Freight Tracker

Real-time freight scheduling visibility across 100+ fulfillment centers, replacing manual spreadsheets.

100+ FCs · 6 countries · real-time vs 30-min lag

React
Apache Flink
DynamoDB
Explore →
2018–2022
Optimization

Reactive Scheduling

Dynamic scheduling engine that optimized freight capacity allocation using demand forecasting and real-time signals.

$800K annual savings · 70% latency reduction

Demand Forecasting
Apache Spark
AWS

The journey

Career at Amazon

Building teams and products across logistics, operations, and AI/ML

Sep 2014 – Mar 2016

Support Engineer

Corporate Logistics · Hyderabad

  • R-Shiny scheduling tool replaced Excel — 60% efficiency gain, 80% error reduction
  • 94% team productivity — 15% above Amazon average
Apr 2016 – Mar 2018

Manager, Web Development & Automation

Corporate Logistics · Hyderabad

  • Heimdall (TAT 6.7→2.2hr), LoFAT ($0.6M saved), 9 serverless products — 5M+ views
  • Automated 42 processes saving 30,000+ man-hours annually
Apr 2018 – Apr 2022

Senior Software Development Manager

Corporate Logistics · Hyderabad

  • Scaled org 3→22 engineers across India, Dubai, Mexico — $43M+ combined impact
  • Delivered Heimdall, DFT, Delay Alert Dashboard across 6 regions
Apr 2022 – Present

Senior Software Development Manager

People Experience & Central Science · Seattle

  • AMG adopted org-wide — 73.2% of 203K internal tables had no descriptions before
  • $1.2M annual savings · 500 hrs/month manual curation eliminated

People first

Building teams, not just products

The work was never just technical. Building the org capable of doing the work was as important as the work itself.

3 → 22

Engineers grown

India · Dubai · Mexico

7

Promotions driven

12-month evidence-based cases

150+

Interviews conducted

Across own team and broader org

4.8 / 5

Connections score

Structured mentorship program

3 countries

Cross-geography

Operated across time zones

VP-level

Executive visibility

3-year roadmaps · VP Monthly Reviews

Hire for judgment, not just skill

In a distributed org, engineers make decisions independently. Interview rubric weighted ambiguous problem-solving over technical trivia.

Promotion from day one

Wrote down what the next level looked like for each engineer from week one. Collected evidence for 12 months — not 3 weeks before the cycle.

Make the problem visible before solving it

The catalog metadata UI was shipped deliberately to surface completeness as a metric — creating the pressure that funded the AMG initiative.

Structured debt, not invisible debt

When shipping with known gaps, documented them in the ORR with owners and dates before going live. Debt registered and paid — not hoped away.

Tools of the trade

Technical Toolkit

Cloud & Infrastructure

AWS Lambda
EC2
S3
ECS
EKS
API Gateway
CloudWatch
IAM
Kinesis
Redshift
DynamoDB
CloudFormation
SNS
KMS

AI / ML

Claude LLM
RAG
Prompt Engineering
Active Learning
Demand Forecasting
Anomaly Detection
ML Productionisation

Data Engineering

Apache Kafka
Apache Flink
Amazon Kinesis
MySQL
PostgreSQL
ETL Pipelines
PII Tagging
Data Obfuscation
Apache Spark
SQL

Frontend & Web

ReactJS
Node.js
TypeScript
R-Shiny

Languages

Python
R

Compliance & Security

GDPR
HIPAA
AWS KMS
Field-level Encryption
Access Control
Penetration Testing
Audit Logging

© 2025 Misbahuddin Mohammed · Senior Engineering Leader