LOADING
AKAnish/kumar
Hire me
Open to ML / AI roles · 2026

python3 — ml_engineer.py

ANISH
KUMAR

View projects Resume
python3 — ml_engineer.py LIVE
Production ML◆Computer Vision◆RAG Pipelines◆FastAPI + Docker◆MLflow◆PyTorch◆vLLM◆Quantization◆TensorRT◆Edge Inference◆Production ML◆Computer Vision◆RAG Pipelines◆FastAPI + Docker◆MLflow◆PyTorch◆vLLM◆Quantization◆TensorRT◆Edge Inference◆
3+
Years building
15+
Projects shipped
<80ms
Inference latency
8+
Tech stack depth
Model Registry

Production-grade ML, end to end.

Computer Vision

Real-Time Object Detection

YOLOv8 on a custom 12K-image dataset, served via FastAPI with TensorRT acceleration. Live webcam demo.

45
Target FPS
12K
Images
WIP
Status
YOLOv8PyTorchTensorRTFastAPI
NLP / LLM

BERT Sentiment Classifier

Fine-tuned BERT for 5-class sentiment. Production REST API handling 500+ RPS, fully Dockerised.

91%
Accuracy
500+
RPS
5
Classes
TransformersONNXFastAPIMLflow
ML Systems

Churn Prediction Pipeline

End-to-end churn predictor with SMOTE, SHAP explainability and a real-time Streamlit dashboard.

87%
Precision
83%
Recall
-18%
Impact
XGBoostSMOTESHAPStreamlit
Computer Vision

FaceNet Attendance

FaceNet embeddings + cosine similarity for automated attendance — 200+ users at 80ms latency.

98.4%
Accuracy
80ms
Latency
200+
Users
TensorFlowFaceNetOpenCVSQLite
NLP / LLM

Document Q&A — RAG

RAG pipeline for multi-document Q&A with FAISS vector search and transformer embeddings.

0.81
BLEU
1K+
Docs
1.2s
Latency
LangChainFAISSHuggingFace
ML Systems

Fraud Detection — XGBoost

Imbalanced classification on 2M credit card records, SMOTE + cost-sensitive training, tracked in MLflow.

99.2%
Recall
0.94
F1
2M
Records
XGBoostSMOTEMLflow
Writing

blog/

Idea in progress

What if a camera could watch the drip so nurses don't have to?

Exploring how computer vision + low-cost edge devices could automate IV bottle monitoring, estimate fluid levels and alert nurses before they run empty.

Read article →
Case study

Document Q&A — building a grounded RAG pipeline

Chunking strategy, embedding choice and FAISS index tuning that pushed BLEU from 0.41 (keyword search) to 0.81 on a 1K+ PDF corpus.

Read article →
The Stack

Models that survive production.

PyTorch · TensorFlow
Training & fine-tuning
YOLO · FaceNet · OpenCV
Computer vision
BERT · LangChain · RAG
NLP & LLMs
FastAPI · ONNX · TensorRT
Inference serving
Docker · MLflow
MLOps & tracking
FAISS · SQL · Pandas
Data & retrieval
Now

$ ls currently_learning/

CUDA
kernel tuning · GPU optimisation
RAGAS
RAG evaluation metrics
vLLM
high-throughput LLM serving
Chroma
vector DB · semantic search
GGUF
quantised model deployment
AK
Anish Kumar
ML Engineer · Open to work
Punjab, India · Remote / Relocate
BTech AI/ML — LPU
Graduating 2027 · 2026 Intern roles open
OPEN_TO_WORK = True
About

I write ML code that survives git push --force.

Final-year BTech AI/ML student building production-ready machine learning systems. I don't just train models — I deploy them. My focus is the full ML lifecycle: data pipelines, training, inference optimisation, API serving and monitoring.

Currently building local LLM inference servers with vLLM, quantised model deployment with GGUF, and RAG evaluation pipelines with RAGAS metrics.

Currently building
  • → Quantised Mistral server
  • → RAG eval with RAGAS
  • → Edge inference w/ GGUF
Training History

Checkpoints.

2024 — 2025
AI Training & Evaluation Specialist
[ Outlier AI ]
RLHF-style evaluation of 500+ LLM responses for safety, reasoning and code quality. Improved prompt acceptance rate by 18%.
2023 — Now
Independent ML Engineer
[ Self-directed / Open source ]
Shipped 6+ end-to-end ML systems across CV, NLP and predictive modelling — Dockerised, API-served, benchmarked.
2023 — 2027
BTech — Artificial Intelligence & ML
[ Lovely Professional University ]
Coursework in Deep Learning, Computer Vision, NLP, Statistics, Data Structures, MLOps.
“Structured feedback that improved prompt acceptance by 18%.”
— Outlier AI · RLHF Evaluation, 500+ responses
Credentials

Certifications.

In progress
Deep Learning Specialization
DeepLearning.AI2026
In progress
MLOps Fundamentals
Google Cloud2026
In progress
Developer Certificate
TensorFlow2026
Completed
Machine Learning by Andrew Ng
Coursera2024
Completed
NLP with Transformers
HuggingFace2025
Completed
Fundamentals of Deep Learning
NVIDIA DLI2025
Follow on GitHub
—
Public repos
—
Total stars
—
Followers
—
Following
GitHub contribution graph
Send a message

LET'S
BUILD.

I reply within 24 hours. Open to ML/AI engineering roles, internships and collaborations.

anish00kumar11@gmail.com
GitHub LinkedIn LeetCode Kaggle