AI & MACHINE LEARNING ENGINEER

Dhruv Parmar.

I turn complex data into intelligent systems.
Built with curiosity. Made for the real world.

Explore my work Download résumé

A little curiosity.
A lot of possibility.

SAN FRANCISCO, CASEARCH · MACHINE LEARNING · SYSTEMSA little about me
01 THE PERSON BEHIND THE CODE

Good models matter.
What they make possible matters more.

I'm Dhruv, an AI/ML engineer focused on search, retrieval, and healthcare intelligence. I work across the full path from a promising idea to a reliable production system.

My work brings together experimentation, careful evaluation, and engineering at scale. I care about finding better evidence, making faster decisions, and building systems people can depend on.

Explore my experience
Daily queries supported
200M
Reduction in p50 search latency
30.2%
Reduction in embedding storage
75%

02 SELECTED WORK

Ideas, put to work.

A selection of production systems and projects
across search, healthcare, and computer vision.

4 selected pieces

PERPLEXITY / SEARCH & RETRIEVAL

Better evidence. Faster answers.

Hybrid retrieval and ranking systems that select relevant evidence for production RAG pipelines.

  • Python
  • PyTorch
  • RAG
Explore the work

My contribution

Combined staged candidate filtering, embedding-based scoring, and selective cross-encoder reranking. Evaluated releases with retrieval benchmarks, LLM graders, human judgments, and A/B tests.

The work supported pipelines serving 200 million daily queries and contributed to 93.0% overall SimpleQA accuracy. Native INT8 quantization reduced embedding storage by 75%, with retrieval quality validated against full-precision baselines.

30.2% lower p50 latency

HCL TECH / HEALTHCARE ML

Data that supports better care.

Healthcare data and ML workflows for patient-risk insights, data quality, and clinical applications.

  • Databricks
  • MLflow
  • FHIR
Explore the work

My contribution

Built Python and SQL pipelines for clinical and claims data, OAuth-secured FHIR integrations, and reusable workflows for feature engineering, model versioning, and monitoring.

Automated validation across more than 6,000 healthcare quality rules. Standardized deployment and governance workflows across client environments, with drift and data-quality monitoring for ongoing reliability.

3 weeks → 3 days to deploy

MASTER'S PROJECT / GAME NODE

Recommendations with a player in mind.

A personalized gaming dashboard that connects Steam activity with useful game recommendations.

  • Steam Web API
  • Recommendations
  • AI
Explore the work

My contribution

Used Steam playtime, achievements, and favorites to personalize recommendations. Added 24-hour caching to reduce redundant API calls and improve response time.

Developed an AI-assisted complaint-resolution workflow that generates contextual recommendations and supports escalation to a human when needed.

24-hour caching for API efficiency

MASTER'S PROJECT / COMPUTER VISION

A new dimension from everyday video.

An end-to-end pipeline that turns monocular smartphone videos into 3D assets for simulation.

  • Python
  • COLMAP
  • Open3D
Explore the work

My contribution

Built a reconstruction workflow with COLMAP and Open3D, from smartphone video to processed point clouds and 3D assets.

Added automated quality control to identify visual and geometric defects, improving the reliability of reconstructed assets for downstream simulation.

Video → simulation-ready 3D

03 EXPERIENCE

From ideas
to impact.

Building at the intersection of machine learning, data, and production engineering.

Get the full résumé
PerplexityMay 2025 — Present

AI/ML Engineer

San Francisco, California

Building retrieval, ranking, and evaluation systems for answer generation at scale.

  • Hybrid retrieval and selective cross-encoder reranking for high-volume search.
  • Offline evaluation, relevance benchmarks, and A/B testing for ranking releases.
  • Distributed model serving and indexing on AWS and Kubernetes.
HCL TechAug 2021 — Jun 2024

Machine Learning Engineer

India

Developing healthcare ML and data platforms for clinical decision workflows.

  • Python and SQL pipelines for clinical, claims, and patient-risk data.
  • Reusable ML workflows with governed features, evaluation, and drift monitoring.
  • Standardized deployments that reduced rollout time from three weeks to three days.

04 EXPERTISE

The tools behind the work.

From a first experiment
to a system that holds up in production.

01 /

Models & retrieval

  • PyTorch
  • Scikit-learn
  • Hybrid retrieval
  • RAG
  • Cross-encoders
  • LLM evaluation
02 /

Data & experimentation

  • Python
  • SQL
  • Airflow
  • Databricks
  • MLflow
  • A/B testing
03 /

Systems & deployment

  • AWS
  • Kubernetes
  • Docker
  • CUDA
  • Rust
  • C++

EDUCATION

M.S. Computer Science

Saint Louis University

B.Sc. Computer Applications

The Maharaja Sayajirao University of Baroda

CERTIFICATIONS

  • AWS Certified AI Practitioner
  • NVIDIA Generative AI LLMs Associate
  • DATABRICKS Machine Learning Associate

05 GET IN TOUCH

Have something
in mind?

I'd love to hear about your team, your idea,
or the problem you're trying to solve.

LET'S START A CONVERSATIONPrefer to start with my résumé?