Brejesh Balakrishnan

Experience

Engineering discipline, applied to ML.

Production software first, machine learning second — in the order that makes ML systems actually hold up.

  1. Jul 2024 – Mar 2026 Bengaluru, India

    Accenture

    Advanced Application Engineering Analyst

    • Optimized AWS-hosted REST APIs (Java / Spring Boot), cutting high-latency responses (>5s) from ~80% to <20% via profiling, SQL query tuning, and caching on a financial-services platform
    • Delivered 200+ story points across Agile/Scrum sprints, translating ambiguous requirements into production features with regular stakeholder demos
    • Maintained >90% automated test coverage (unit, contract, BDD) tracked via SonarQube
    • strengthened CI/CD reliability through code reviews and linting/test gates
    • Built and modified REST endpoints with business-rule logic
    • debugged DevOps pipeline issues and optimized release workflows on AWS (Jenkins CI/CD)

    Java · Spring Boot · AWS · REST APIs · SQL · Jenkins · SonarQube · Git · Agile/Scrum

  2. Jul 2023 – Aug 2023 Chennai, India

    UnisLink

    Data Engineering Intern

    • Researched and benchmarked natural-language-to-SQL (NL-to-SQL) approaches, documenting accuracy–complexity tradeoffs to guide early product direction
    • Built and validated SQL Server replication/synchronization checks and a replication test plan to improve data consistency across large tables

    SQL Server · NL-to-SQL Research · Data Replication · Documentation

  3. Jul 2022 – Sep 2022 Tiruchirappalli, India

    Yaltech Global Consulting

    Machine Learning Intern

    • Trained and cross-validated classification/regression models (Logistic Regression, Random Forest) in Python/scikit-learn on structured datasets
    • Performed EDA, data cleaning, and feature engineering (Pandas/NumPy)
    • communicated findings to non-technical stakeholders with clear visualizations

    Python · Pandas · NumPy · scikit-learn · EDA · Cross-Validation

  4. Jan 2022 – Mar 2022 Chennai, India

    Youth India Foundation

    Sponsorship Intern

    • Built and maintained a structured database of potential sponsors
    • supported outreach and follow-ups
    • Assisted in planning and executing a coding event, strengthening stakeholder communication and delivery under deadlines

    Stakeholder Management · Outreach · Event Coordination

  5. Apr 2021 – Jun 2021 Tiruchirappalli, India

    TREC-STEP

    Incubation & Proposals Intern

    • Supported senior leadership on international funding proposals, including an EU Horizon 2020 submission on Responsible AI (RAI)
    • Coordinated incubation activities — site visits, documentation, and ecosystem-building program execution

    Proposal Research · Responsible AI · Documentation · Program Support

Impact — measured, not claimed

0+

Story Points Delivered

Agile delivery on a financial-services platform at Accenture

0% → <20%

High-Latency Responses Cut

AWS REST APIs fixed via profiling, SQL tuning, and caching

0%+

Automated Test Coverage

Unit, contract, and BDD tests tracked via SonarQube

0+

Tests Across Shipped ML Systems

343 in the RAG backend + ~399 in the churn MLOps stack

0%

Docker Image Reduction

11.4 GB → 4.9 GB via CPU-only torch and a split slim UI image

0%

Macro-F1 Gain (SageMaker HPO)

Bayesian tuning with 2-instance distributed spot training

PR-AUC 0.62 → 0.67

Churn Model Lift (Optuna)

With isotonic calibration and a cost-based decision threshold

0.72 → 0.90

LLM Faithfulness Improvement

Two-tier grounding eval caught and fixed a real bug (n=50, seeded)

recall@k · MRR · nDCG

Retrieval Eval Discipline

Deterministic harness over a hand-labeled golden set — decisions on data, not vibes