Machine Learning Engineer
Summary
Hands-on Machine Learning Engineer in Hyderabad who owns end-to-end applied ML solutions for Operations — audio/video and computer vision work (transcription QA, detection, tracking, redaction) using OpenCV, plus LLM-based agentic AI workflows, covering evaluation, deployment, monitoring, and reusable self-serve tooling.
Machine Learning Engineer
Location: Hyderabad
Key Skills: Applied ML | Audio, Video & Computer Vision | Agentic AI
About the Role
We are looking for a hands-on Machine Learning Engineer to turn operational problems into reliable, scalable ML solutions. You will combine a strong foundation in statistics with practical experience in audio, video, computer vision, and LLM applications. You will own solutions from problem definition and data exploration through evaluation, deployment, monitoring, and continuous improvement.
The role requires sound engineering judgment, clear stakeholder communication, and the ability to build reusable platforms and self-serve tools that improve quality, reduce manual effort, and deliver measurable business value.
Key Focus:
- Strong hands-on Machine Learning experience with a solid foundation in Statistics.
- Practical experience applying ML to Audio, Video and Computer Vision problems using OpenCV.
- Ability to design and build end-to-end ML solutions independently.
- Experience defining validation strategies, quality checks and evaluating model performance.
- Hands-on experience developing Agentic AI workflows.
- Strong ownership mindset with the ability to independently drive deliverables and take accountability for outcomes.
What You Will Do
- Own problem discovery and prioritization: Translate Operations requests into clear requirements covering sample inputs and outputs, data volume, quality targets, deadlines, budget, and human review needs. Prioritize by business impact, urgency, effort, and dependencies; agree on delivery milestones and service expectations.
- Build end-to-end ML solutions: Develop data ingestion, preprocessing, inference, validation, review, and output workflows. Choose practical baselines, benchmark on representative samples, and make informed build-versus-buy decisions.
- Develop audio and video applications: Apply ML to tasks such as transcription quality assessment, object detection, tracking, and sensitive-content redaction. Use OpenCV and appropriate models to handle challenging conditions such as motion blur, occlusion, low resolution, and varied scene density.
- Design meaningful quality checks: Validate input schemas, data completeness, annotation consistency, and output integrity. Build evaluation datasets, analyze errors by scenario, and define acceptance thresholds and human-in-the-loop escalation paths.
- Scale efficiently: Estimate throughput, turnaround time, infrastructure needs, and cost per unit. Optimize batching, CPU/GPU usage, parallel processing, and detection/tracking strategies while measuring their effect on quality.
- Build reusable automation and agentic workflows: Create configuration-driven components, LLM evaluation frameworks, and tools for Operations. Develop agents with tool integration, bounded actions, validation, retries, and human approval where needed.
- Operate and improve production systems: Implement logging, monitoring, versioning, failure recovery, and documentation. Track request trends, adoption, quality, reliability, and engineering effort to identify opportunities for standardization and self-service.
What Success Looks Like
You deliver solutions with agreed quality, turnaround time, and cost targets; demonstrate improvements through representative evaluation and error analysis; and build reusable capabilities that reduce recurring engineering work. Operations teams can use and support your tools with clear documentation and appropriate human oversight.
Skills
Sign in to read and join the discussion.