Lead AI / ML Engineer
Building ML/AI systems that move from research to production and scale.
I help teams turn AI ideas into reliable, high-impact products with measurable outcomes.
- Distributed training for SLMs, LLMs and LVLMs
- High-throughput, cost-efficient inference at scale
- Data-centric AI pipelines with production-grade observability
- Production-ready agentic AI and evaluation workflows
Flagship Open Source Product
harneXa/nexa-gauge
A graph-based evaluation toolkit for LLM and RAG systems with repeatable quality checks, upfront cost visibility, and clean per-case outputs for analysis.
- Graph-native evaluation flow (scan -> claims -> metrics -> eval)
- Cost visibility before runtime with estimate-first execution
- Cache-aware runs to avoid duplicate spend and recomputation
- Coverage across relevance, grounding, redteam, GEval, and reference scoring
- Production-friendly CLI for run, estimate, and cache management
- Scales with control across utility and metric nodes
harneXa/nexa-prism
Coming Soon10+
Years in ML & AI Research and Engineering
2–3x
Inference Performance Gains
~50%
ML Infra Cost Reduction
Published
ICCV and AAMAS
Weeks → Hours
Governed pipelines and data workflows that accelerate model iteration
6–10%
Use-case specific model gains through architecture tradeoffs and training pipeline design
Capabilities
What I Work With
Weighted from public GitHub activity with recency, stars, and topic signals.
Projects
Personal Work
A glimpse into how I spend my personal time building, exploring, and learning.
harneXa/nexa-gauge
ReleasedA graph-based evaluation toolkit for LLM and RAG systems with repeatable quality checks, upfront cost visibility, cache for reusability and clean per-case outputs for analysis. Metric suport: Grounding, Relevance, RedTeam, Geval, Reference-based.
Self-Driving Vehicle
AccomplishedPerception and control modules for autonomous vehicles built on the Udacity SDC curriculum. Covers lane detection, traffic sign classification, behavioral cloning, LIDAR/RADAR sensor fusion via Extended Kalman Filter, jerk-minimizing path planning, and a PID controller for steering and throttle.
Deep Reinforcement Learning
AccomplishedCore Deep RL algorithms implemented across Unity ML-Agents environments. Covers DQN and Double-DQN for discrete action spaces, REINFORCE for Atari Pong, DDPG for continuous robotic arm control, and Multi-Agent DDPG for a collaborative/competitive multi-agent setting.
Writing
Latest Posts
Recent articles on practical AI engineering and production model workflows.