Algorithm Developer - Matlab - Bangalore

Vidpro Consultancy Services; Bengaluru; On-site; INR 900,000 - 1,300,000; Full time

Job summary

Vidpro Conslutancy is hiring an Algorithm Developer / Engineer in Bangalore with 3+ years of hands-on experience to design, develop and validate data fusion, tracking and recognition algorithms. You will prototype in MATLAB and hand the benchmarked code to the software team for embedded use.

Key skills include MATLAB and Python programming, matrix mathematics, estimation theory and Kalman filter variants. You will translate equations into code and build simulations for rigorous testing on site.

Qualifications

• 3+ years developing algorithms in MATLAB or Python.
• Design and validate data fusion, tracking and recognition algorithms.
• Strong MATLAB and Python programming skills.

Responsibilities

• Convert equations into working code for fusion/tracking/recognition algorithms.
• Design and implement fusion, tracking or recognition algorithms.
• Build sensor models from interface specs.
• Create simulations and generate test data to validate algorithms.
• Measure performance with accuracy, track continuity, false track rate and latency.
• Benchmark algorithms and recommend best performer.
• Define test cases and perform standalone tests.
• Write clean, documented code and explain the equations, design and results.
• Present designs at design reviews and address action points.
• Support C/C++ implementation by software team.

Skills

MATLAB

Python

Matrix math

Probability & stats

Estimation theory

Numerical methods

Equations to code

C basics

Independent coding

Education

B.E./B.Tech or M.Sc in Math/Physics

M.E./M.Tech preferred

Job description

*Greetings from Vidpro Conslutancy.*

We are hiring for Algorithm Developer / Engineer with 3+ years of experience for once of our Aerospace and defence Company in Bangalore

• Location: Bangalore, full-time at the client site. Remote or hybrid work is not possible.

Role summary

Design, develop and validate data fusion, tracking and recognition algorithms. Prototype them in MATLAB measure their performance, and hand the benchmarked algorithms to the software team for embedded implementation.

Key responsibilities

• Convert mathematical equations and algorithm descriptions into working code.
• Design and implement fusion, tracking or recognition algorithms.
• Build sensor models from interface specifications.
• Build simulations and generate test data to check the algorithms.
• Measure performance with numbers: accuracy, track continuity, false track rate and latency.
• Benchmark candidate algorithms and recommend the best performer.
• Define test cases and carry out standalone algorithm testing.
• Write clear, commented code and document the equations, design and results.
• Present designs at formal design reviews and close the action points raised.
• Support the software team when the algorithm is converted to C/C++.

Qualification

• B.E./B.Tech in ECE, EEE, CSE or Aerospace, or M.Sc in Mathematics or Physics. M.E./M.Tech is preferred.

Minimum 3 years of hands-on algorithm development in MATLAB or Python.

*Must-have skills for every candidate*

Area What you must be able to do

MATLAB programming

(MATLAB profile)

Scripts and functions, matrix indexing, vectorised code, structures and cell arrays, plotting, file handling and debugging. Write algorithms with core MATLAB functions, without depending on toolboxes

Python programming

Data types, functions, classes, lists and dictionaries, file and exception handling, and debugging. NumPy arrays, slicing and broadcasting, SciPy linear algebra and Matplotlib

Matrix mathematics

Multiplication, transpose, inverse, determinant and rank. Eigenvalues and eigenvectors. Solving linear equations. Positive-definite matrices. Cholesky, QR and SVD decompositions

Probability and statistics

Random variables, mean, variance, covariance and correlation. Gaussian distribution. Bayes rule. Generating and analysing noise

Estimation theory

Least squares, and the Kalman filter family (KF, EKF, UKF): predict step, update step, process noise and measurement noise. Coordinate frames Polar, spherical and Cartesian conversion. NED and ECEF frames. 2D and 3D rotation matrices

Numerical methods Interpolation

numerical integration and differentiation, root finding, and awareness of precision errors

Equations to code

Read a derivation, implement it correctly, and check the output against a hand calculation

C basics

Data types, loops, functions, arrays, pointers and structures. Read and write simple C programs

Working independently

Write code without internet search or AI tools, recalling common library functions and syntax from memory

*Algorithm areas*

State estimation and tracking

EKF, UKF, IMM filters, motion models (constant velocity, constant acceleration, coordinated turn), noise and bias modelling, track-to-track fusion, covariance intersection

Data association and track management

Gating, GNN, JPDA, MHT, assignment algorithms (Hungarian, auction), track initiation, confirmation and deletion, duplicate track merging

Sensor resource management

Task scheduling, rule-based priority logic, optimisation, information-gain based sensor tasking

Target detection and recognition (Python required)

OpenCV, PyTorch or TensorFlow, CNN and transformer-based object detectors, IR and EO image processing, video object tracking, dataset preparation and augmentation, model export to ONNX