Bilal Saleem
I am an AI and software engineer with more than four years of experience building production systems across startup and enterprise environments.
About
I am an AI and software engineer with more than four years of experience building production systems across startup and enterprise environments. As a founding engineer of a multilingual voice AI startup, I have developed Python and FastAPI backends, integrated leading language models, implemented tool-calling workflows, and deployed reliable services on AWS. I enjoy turning emerging AI capabilities into practical products with strong testing, observability, and operational controls.
My experience also includes distributed speech-model training, NLP research, government technology projects, code-migration automation, and production C++ systems. I bring a hands-on approach to asynchronous workflows, evaluation pipelines, cloud infrastructure, and stakeholder collaboration, and I am comfortable working from early prototypes through deployment and performance optimisation.
- 6 years' experience
- Joined October 2026
Industry experience
- Software & Internet
- Computers & Electronics
- Government
- Media & Entertainment
- Transportation & Logistics
- Education
Work experience· 4 roles
Built the backend for a multilingual voice assistant using Python and FastAPI, integrating OpenAI, Anthropic, and Cohere models with application tools, native function calling, asynchronous workflows, and AWS ECS deployment. Engineered long-running AI workflows with SQS-backed jobs, signed webhooks, retries, failure handling, concurrency controls, and usage metering. Developed reliability and observability capabilities including circuit breakers, health checks, Prometheus metrics, latency instrumentation, automated testing, CI, and load testing. Built a speech-data and training pipeline for more than 600,000 samples using WebDataset and S3, supporting distributed multi-GPU Whisper large-v3 fine-tuning with PEFT/LoRA and evaluation through W&B and MLflow.
Built an NLP classification pipeline using Stanford CoreNLP to analyse media-bias patterns across film scripts. Also delivered a prototype automation tool for migrating code from RTC to GitHub.
Delivered proof-of-concept systems for government clients, contributing to technical implementation, performance testing, and deployment into Red Hat OpenShift environments.
Education
Hire Bilal Saleem today
Post a job and invite Bilal, or contact them directly. No hidden commission.